Citation
Wireless Biomedical Telemetry Systems Based on Flexible Metamaterial Circuits and Advanced Rf Architectures

Material Information

Title:
Wireless Biomedical Telemetry Systems Based on Flexible Metamaterial Circuits and Advanced Rf Architectures
Creator:
Cheng, Xiaoyu
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (165 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Electrical and Computer Engineering
Committee Chair:
YOON,YONG KYU
Committee Co-Chair:
EISENSTADT,WILLIAM R
Committee Members:
LIN,JENSHAN
WONG,FONG
KIM,GLORIA JUNG A
Graduation Date:
12/13/2013

Subjects

Subjects / Keywords:
Amplifiers ( jstor )
Antenna design ( jstor )
Antennas ( jstor )
Bruxism ( jstor )
Electric potential ( jstor )
Endoscopes ( jstor )
Patch antennas ( jstor )
Permittivity ( jstor )
Sensors ( jstor )
Transmission lines ( jstor )
Electrical and Computer Engineering -- Dissertations, Academic -- UF
amplifiers -- antennas -- biomedical -- bluetooth -- bruxism -- endoscope -- metamaterial -- mhealth -- patch
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Electrical and Computer Engineering thesis, Ph.D.

Notes

Abstract:
Exploration of several novel wireless components and their applications in different bio-medical telemetry systems are presented.Individual RF components including low noise amplifier and several different types of flexible antennas based on liquid crystalline polymer (LCP) are proposed,designed, fabricated and characterized. All those components are then employed in two platforms including a novel wireless capsule endoscope and a wireless mouth guard. Due to the parasitics of the IC packaging,amplifiers always show gain roll off. The amplifier gain is higher at low frequency, but is lower at high frequency. A frequency equalizer is used to compensate the gain roll off. With the area of 6mm×8mm, a gain flatness of ±1.1dB is demonstrated from 100 MHz to 7 GHz with a nominal gain of 33.5dB. Antennas for implants including capsule endoscope applications are difficult to implement due to the dispersive and lossy nature of human body. With the assistance of the human body model from Ansys Inc., a family of new wrappable antennas based on flexible substrates are designed,fabricated and characterized. Those wrappable antennas are based on patches,and are wrapped into a cylindrical capsule shape. The antenna size is minimized by using inductive slot metamaterial particles and electrical vias loading.Those antennas show a good omni-directional pattern while offering electromagnetic interference (EMI) shielding and physical protection to the capsule endoscope systems. Bruxism involves the activity of grinding and clenching teeth for non-function purposes. While traditional wired systems constraint patients from moving freely, a new wireless mouth guard is developed based on a capacitance to digital signal converter. However, this system suffers from common mode noise generated by oral tissue moving. A time constant method with quasi-differential input is used on a new mouth guard system and common mode noise is suppressed using algorithms in the microprocessor. Metamaterials are a family of artificial structures which demonstrate behaviors that are not readily available in the nature. They are good candidates for sensing due to their tunability. Based on this feature, a battery free and wireless mouth guard for Bruxism is also developed and demonstrated. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: YOON,YONG KYU.
Local:
Co-adviser: EISENSTADT,WILLIAM R.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-12-31
Statement of Responsibility:
by Xiaoyu Cheng.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Embargo Date:
12/31/2014
Resource Identifier:
907646385 ( OCLC )
Classification:
LD1780 2013 ( lcc )

Downloads

This item has the following downloads:


Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EZ8YFOBH1_YG9Q12 INGEST_TIME 2014-05-13T17:06:51Z PACKAGE UFE0045407_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

1 WIRELESS BIOMEDICAL TELEMETRY SYSTEMS BASED ON FLEXIBLE METAMATERIAL CIRCUITS AND ADVANCED RF ARCHITECTURES By XIAOYU CHENG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Xiaoyu Cheng

PAGE 3

3 To my parents

PAGE 4

4 ACKNOWLEDGEMENTS This work was supported in part by the National Science Foun dation (NFS) under funding ECCS 1132413. First of all, I would like to thank my graduation committee chair, Dr. Yong Kyu Yoon for his continuous supports and patience. I would also like to thank the rest of my graduation committee members: Dr. Fong Wong, Dr. Jenshan Lin, Dr. Gloria J. Kim and Dr. William Eisenstadt for all their valuable suggestions, discussions and their time and patience to serve as my graduation committee members. This work would not be possible without the support from my research group members, including all my colleagues at the Multidisciplinary nano and Microsystems (MnM) Laboratory and the Interdisciplinary Microsystems Group (IMG). Especially, I would acknowledge Dr. David E. Senior, Dr. Jun g kwun Kim, Pitfee Jao, Cheolbok Kim, Jun Shi, Melroy Machado and Dr. Kyo u ng Tae Kim for all their assistance and time.

PAGE 5

5 TABLE OF CONTENTS Page ACKNOWLEDGEMENTS ................................ ................................ ................................ ............. 4 LIST OF TABLES ................................ ................................ ................................ ........................... 7 LIST OF FIGURES ................................ ................................ ................................ ......................... 8 LIST OF ABBREVIATIONS ................................ ................................ ................................ ........ 13 ABSTRACT ................................ ................................ ................................ ................................ ... 15 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 18 1.1 Motivation ................................ ................................ ................................ ........................ 18 1.2 Microwave I nteraction with H uman B ody ................................ ................................ ...... 23 1.3 Flexible E lectronics ................................ ................................ ................................ ......... 28 1.4 Metamaterial and N on planar 3D A rchitect ures for B iomedical A pplications ............... 31 1.5 Objective and T hesis S tructure ................................ ................................ ........................ 32 2 METAMATERIAL CIRCUITS ................................ ................................ ............................. 40 2.1 Background and Theory ................................ ................................ ................................ .. 40 2.2 Split Ring Resonator (SRR) and Complimentary Split Ring Resonator (CSRR) ........... 41 2.3 Metamaterial Applications ................................ ................................ ............................... 44 2.3.1 As a Homogeneous M aterial ................................ ................................ ............... 44 2.3.2 Loading onto Hosting Transmission L ines ................................ .......................... 44 2.3.3 Loaded into Hosting Bulky W aveguide ................................ .............................. 45 2.3.4 SRR Based Tunable D evice ................................ ................................ ................ 47 2.3.5 SRR Itself as an Electrically Small and Tunable A ntenna ................................ .. 49 2.3.6 Other A pplications ................................ ................................ ............................... 50 3 RF SUBSYSTEMS ................................ ................................ ................................ ................. 59 3.1 Microwave Board Level Design Considerations ................................ ............................. 59 3.2 RF Transceiver ................................ ................................ ................................ ................. 66 3.3 Amplif ier ................................ ................................ ................................ .......................... 68 4 ANTENNAS FOR BIOMEDICAL APPLICATIONS ................................ .......................... 81 4.1 Micristrip Patch Antenna : Design and Applications ................................ ...................... 81 4.2 Wrapped Patch Antenna with Omni directional Radiation Pattern ................................ 82 4.3 Compact Wrapped Patch Antennas ................................ ................................ ................. 84 4.3.1 By Loading Shorting V ias ................................ ................................ ................... 84

PAGE 6

6 4.3.2 By Metamaterial Loading on the Back G round ................................ .................. 85 4.3.3 By M et amaterial Loading on the Patch S urface ................................ .................. 87 4.4 Multifunctional Wrapped Helix Antenna ................................ ................................ ........ 89 4.5 Super Compact Wrapped Patch Antenna with Inductive Slot Loading .......................... 91 5 WIRELESS CAPSULE ENDOSCOPE ................................ ................................ ............... 108 5.1 System Scopes and Ra dio Specifications ................................ ................................ ...... 108 5.2 Planar Patch Antenna with Inductive Loading ................................ .............................. 110 5.3 Capsule Circuitry Design and Antenna Integration ................................ ....................... 111 5.4 System Controller and Communication Protocol Design ................................ .............. 112 5.5 Hybrid Circuitry Design ................................ ................................ ................................ 112 5.6 Sensor and Camera Implementation ................................ ................................ .............. 113 6 SMART MOUTH GUARD ................................ ................................ ................................ .. 126 6.1 Syst em Requirements ................................ ................................ ................................ .... 126 6.2 Wireless Mouth Guard Based on Capacitive Force Sensor ................................ ........... 128 6.2.1 Sensor Principle ................................ ................................ ................................ 128 6.2. 2 Sensor Fabrication ................................ ................................ ............................. 131 6.2.3 Sensor Electronics ................................ ................................ ............................. 132 6.3 System Characterization ................................ ................................ ................................ 133 6.4 Blue Guard: A Bluetooth Enabled mHealth System for Bruxism Management ........... 134 6.5 A Wireless and Battery Free Mouth Guard for Bruxism Management Based on Flexible Metamaterial Particles ................................ ................................ .................... 138 6.6 Future Work ................................ ................................ ................................ ................... 142 7 CONCLUSION ................................ ................................ ................................ ..................... 156 LIST OF REFERENCES ................................ ................................ ................................ ............. 160 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 165

PAGE 7

7 LIST OF TABLES Table Page 1 1 Microwave characteri stics of human body and head ................................ ......................... 33 1 2 Antenna parameter comparison: In free space and beside the human head model .......... 33 3 1 Bill of material (BOM) of the cascaded amplifier ................................ ............................ 75 4 1 Composition of a human body phantom solution. ................................ ............................. 97

PAGE 8

8 LIST O F FIGURES Figure Page 1 1 4mm accuracy male human body model from Ansys ................................ ........................ 34 1 2 2mm accuracy human body models from Ansys ................................ ............................... 35 1 3 2mm accuracy human body head model from Ansys ................................ ........................ 35 1 4 Full accuracy human body model from Ansys ................................ ................................ .. 36 1 5 Dipole antenna with 4mm human body head model ................................ ......................... 36 1 6 Simulated 3D radiation pattern ................................ ................................ .......................... 37 1 7 SAR simulation based on the human body mode l ................................ ............................. 37 1 8 Path loss at different frequency bands in human body ................................ ...................... 38 1 9 Structure of this thesis work ................................ ................................ .............................. 39 2 1 Materials of different kinds ................................ ................................ ................................ 51 2 2 Metamaterial with simultaneous negative permittivity and permeability .......................... 51 2 3 Left handed transmission line ................................ ................................ ............................ 51 2 4 CRLH Model ................................ ................................ ................................ ..................... 52 2 5 Split r ing r esonator (SRR) and its equivalent circuits ................................ ....................... 52 2 6 Double SRR and its equivalent circuits ................................ ................................ ............. 52 2 7 Complementary s plit r ing r esonator (CSRR) and its equivalent circuits ........................... 53 2 8 Negative refrac tion index ................................ ................................ ................................ ... 53 2 9 Simulated wave propagation through a material with negative refraction index .............. 54 2 10 Microstrip loaded with SRR arrays ................................ ................................ .................... 54 2 11 S parameter of SRR loaded microstrip ................................ ................................ .............. 54 2 12 Hollow w aveguide l oaded with SRR a rray ................................ ................................ ........ 55 2 13 Transmission c oefficients (S21) of w aveguide l oaded with SRR a rray ............................ 55 2 14 Waveguide l oaded with SRR ................................ ................................ ............................. 55

PAGE 9

9 2 15 Waveguide c onfiguration ................................ ................................ ................................ ... 56 2 16 Waveguide loaded with SRR for in situ tuning ................................ ................................ 56 2 17 Measured in situ t u nability of the pro posed waveguide ................................ .................... 57 2 18 Transmission at 6GHz with respect to different misalignment level ................................ 57 2 19 Proposed electrically small and tunable antenna ................................ ............................... 57 2 20 Simulated return loss of the SRR a ntenna ................................ ................................ ......... 58 3 1 Different transmission lines ................................ ................................ ............................... 75 3 2 Lumped element model of a tra nsmission line ................................ ................................ .. 75 3 3 Different microstrip bend ................................ ................................ ................................ ... 75 3 4 Superheterodyne r eceiver ................................ ................................ ................................ ... 76 3 5 Intermodulation ................................ ................................ ................................ .................. 76 3 6 Amplifier power sweep response ................................ ................................ ....................... 76 3 7 Single stage amplifier and schematic ................................ ................................ ................. 77 3 8 Measured single stage amplifier performance ................................ ................................ ... 77 3 9 Cascaded amplifier ................................ ................................ ................................ ............. 78 3 10 Frequency response of the cascaded amplifier ................................ ................................ .. 78 3 11 Frequency leveling network: s chematic and layout ................................ ........................... 79 3 12 Frequency leveling network performance ................................ ................................ ......... 79 3 13 Desired frequency response of the frequency equalizer ................................ .................... 79 3 14 Comparison of the desired response and simulated response of the frequency equalizer ................................ ................................ ................................ ............................. 80 3 15 Fabricated broadband amplifier with the frequency equalizer ................................ .......... 80 3 16 Measured amplifier gain with the frequency equalizer ................................ ...................... 80 4 1 Microstrip patch antenna schematic ................................ ................................ ................... 97 4 2 Field d istribution on a patch antenna ................................ ................................ ................. 98 4 3 Patch antennas ................................ ................................ ................................ .................... 99

PAGE 10

10 4 4 Conventional and folded patch antenna performance ................................ ........................ 99 4 5 Folded patch antenna in a rectangular waveguide shape ................................ ................... 99 4 6 Performance of a folded patch antenna with PCB filling ................................ ................ 100 4 7 Compact omnidirectional folded patch antenna ................................ .............................. 100 4 8 Measured and simulated return loss ................................ ................................ ................. 100 4 9 Radiation pattern ................................ ................................ ................................ .............. 101 4 10 Folded patch antenna with CSRR loaded on the ground plane ................................ ....... 101 4 11 Return loss ................................ ................................ ................................ ....................... 101 4 12 Fabr icated antenna ................................ ................................ ................................ ........... 102 4 13 Patch antenna folded in a cylindrical shape ................................ ................................ ..... 102 4 14 CSRR loaded patch antenna ................................ ................................ ............................. 102 4 15 Fabricated antenna and measured results ................................ ................................ ......... 103 4 16 Schematic of the proposed helix antenna ................................ ................................ ......... 103 4 17 Fabr i cated a ntenna ................................ ................................ ................................ ........... 103 4 18 Antenna far field performance ................................ ................................ ......................... 104 4 19 Near field charging performance ................................ ................................ ..................... 104 4 20 Frequency dependent human body parameters ................................ ................................ 104 4 21 Inductive notch loading on a microstrip component ................................ ........................ 105 4 22 Wrapped patch antenna without notch loading ................................ ................................ 105 4 23 Wrappable patch antenna with inductive loading ................................ ............................ 105 4 24 Fabricated patch antenna with inductive loading on a flexible substrate ........................ 106 4 25 Return loss of the fabricated antenna under different conditions ................................ .... 106 4 26 Antenna radiation pattern ................................ ................................ ................................ 107 4 27 3D radiation pattern with human body model ................................ ................................ 107 4 28 Radiation pattern change according to arm positions ................................ ...................... 107

PAGE 11

11 5 1 2.4 GHz wrapped patch antenna with inductive loading ................................ ................ 117 5 2 Measured return loss ................................ ................................ ................................ ....... 117 5 3 Simulated antenna radiation pattern ................................ ................................ ................ 117 5 4 Proposed wrappable capsule endoscope system ................................ ............................. 118 5 5 Wireless endoscope based on hybrid circuits. ................................ ................................ 120 5 6 Temperature characterization of the capsule endoscope ................................ ................ 120 5 7 Capsule endoscope block diagram ................................ ................................ .................. 123 5 8 Camera imaging distance optimization: ................................ ................................ ........... 124 5 9 Wireless data collector for the wireless endoscope ................................ ......................... 125 6 1 Capacitance Sensors ................................ ................................ ................................ ........ 143 6 2 Different kinds of parallel capacitive sensors ................................ ................................ 143 6 3 Fabricated flexible biting force sensors ................................ ................................ ........... 144 6 4 Capacitive sensor characterization, PDMS thickness i s 3 mm ................................ ........ 144 6 5 System block diagram ................................ ................................ ................................ ..... 145 6 6 RF Balun on the system ................................ ................................ ................................ .. 145 6 7 Fabricated system ................................ ................................ ................................ ............ 147 6 8 Mouth guard with circuitry integrated ................................ ................................ ............ 147 6 9 Biting characterization ................................ ................................ ................................ .... 147 6 10 Teeth clenching at different positions ................................ ................................ ............. 148 6 11 Block diagram of the Bruxism management mHealth system ................................ ......... 148 6 12 Capacitive sensor for Bruxism monitoring ................................ ................................ ..... 148 6 13 RC network for Bruxism monitoring ................................ ................................ ............... 149 6 14 Quasi differential input ................................ ................................ ................................ ... 149 6 15 Wireless Bruxism management system ................................ ................................ .......... 150 6 16 Android application for Bruxism management: ................................ .............................. 152

PAGE 12

12 6 17 Measured output as a function of applied force on sensors error bars show the repeatability ................................ ................................ ................................ ...................... 153 6 18 Varactor loaded split ring resonator ................................ ................................ ................. 153 6 19 Piezoel ectric film characteristics ................................ ................................ .................... 153 6 20 Varactor tunability ................................ ................................ ................................ .......... 154 6 21 Varactor loaded tunable SRR ................................ ................................ ........................... 154 6 22 Assembled wireless battery free mouth guard ................................ ................................ 154 6 23 Measurement setup ................................ ................................ ................................ ......... 154 6 24 Measured frequency response at different weight levels ................................ ................ 155 6 25 Resonance frequency shift as a function of the applied weight ................................ ...... 155

PAGE 13

13 L IST O F ABBREVIATIONS BER Bit Error Rate BOM Bill of Material CDMA Code Division Multiple Access CRLH Composite Right and Left Handed CSRR Compl e mentary Split Ring Resonator CTE Coefficient of Thermal Expansion ECG E lectrocardiogram EDGE E nhanced D ata rate for GSM E volution EM Electromagnetic EMC Electromagnetic Compatibility EMG E lectromyography EMI E lectromagnetic I nterference FCC Federal Communications Commission FDA Food and Drug Administration FDTD Finite D ifference T ime D omain FEM Finite Element Metho d FHSS F requency H opping S pectrum S preading FSK F requency S hift K eying GDP Gross Domestic Product GMSK Gaussian Minimum S hift Keying GSM Global System for Mobile Communications HFSS High Frequency Structure Simulator IF I ntermediate F requency ISM The I ndustrial, S cientific and M edical band

PAGE 14

14 IP3 3rd o rder I ntermodulation P oint I TU International Telecommunication Union LCP L iquid C rystalline P olymer LH Left Handed MIC Microwave Integrated Circuit MICS Medical Implant Communication Service MoM Method of Moments PA Power Amplifier PCB Print Circuit Board PDMS Polydimethylsiloxane PTFE Polytetrafluoroethylene QPSK Quadrature Phase Shift Keying RF Radio Frequency RH Right Handed SAR Specific Absorption Rate SMT Surface Mount SNR Signal to Noise Ratio SRR Split Ring Resonator TEM T ransverse E lectromagnetic VNA Vector Network Analyzer VSWR Voltage Standing Wave Ratio WHO World Health Organization

PAGE 15

15 Abstract of D issertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy WIRELESS BIOMEDICAL TELEMETRY SYSTEMS BASED ON FLEXIBLE METAMATERIAL CIRCUITS AND ADVANCED RF ARCHITECTURES By Xiaoyu Cheng December 2013 Chair: Yong Kyu Yoon Major: Electrical and Computer Engineering Exploration of several novel wireless components and their application s in different biomedical telemetry systems are presented. Individual RF components including a low noise amplifier and several different types of flexible antennas based on liquid crystalline polymer (LCP) are proposed, designed, fab ricated and characterized. All those components are then employed in two platforms such as a novel wireless capsule endoscope and a wireless mouth guard. Due to the parasitics of the IC packaging, amplifiers always show gain roll off. The amplifier gain is higher at low frequency, but is lower at high frequency. A frequency equalizer is used to compensate the gain roll off. One w ith an area of 6 mm 8 mm, a gain flatness of 1.1 dB is demonstrated from 100MHz to 7GHz with a nominal gain of 33.5dB. Antenna s for biomedical implants including capsule endoscope applications are difficult to implement due to the dispersive and lossy nature of the human body. Bas ed on a conventional patch antenna, we first developed a folded patch antenna buil t on a Rogers Duroid 5880 rigid substrate, targeting 2 4GHz band T h e antenna has the same antenna length of 40 mm (z direction) and the resultant waveguide structure has a d imension of 20 mm x 20 mm in its cross section, demonstrating a much reduc ed footprint. We then developed a new generation of the

PAGE 16

16 folded patch antenna with via loading and metamaterial particle such as complementary split ring resonator (CSRR) loading. Com paring to th e original folded patch antenna the metamaterial loaded patch 74% size reduction is achieved after CSRR loading. With the assistance of the human body model from Ansys Inc., a new kind of wrappable antennas based on flexible substrate s are designed, fabricated and characterized. Those wrappable antennas are designed on a liquid crystalline polymer (LCP) flexible substrate with a thickness of 4 mils (0.1 mm) and a dielectric constant of 2.9. With a patch width of 18.5 mm and a capsule diameter of 10 mm, the resonant radiation frequency of the antenna is as low as 433MHz. Based on the concept of wrappable electronics, a new kind of low cost and high performance wirel ess endoscope which carries multiple sensors and a camera is developed. Bruxism involves the activity of grinding and clenching teeth for non function purposes. While traditional wired systems constraint patients from moving freely, a new wireless mouth gu ard is developed based on capacitance to digital signal converter This system has sensitivity as low as 4fF with the maximum sensing freque ncy of 90Hz. The system operates at 2.4GHz frequency band and is powered by CR3240 coin cell. A time constant method with quasi differential input is used on a new mouth guard system and common mode noise is suppressed using algorithms in the microprocessor. In this version, a capacitive sensor with inherent capacitance 10pF is used, the capacitor is charged and dischar ged through a 6 mega ohm resistor. A processor measures the discharging time. Biting changes the capacitance, and hence the discharging time The processor is programmed to digitize the value and provide the output to a smart phone via the Bluetooth Low En ergy wireless

PAGE 17

17 channel. Comparing to an old version of Bluetooth system which consumes 40mA of the peak current, our system consumes 20mA peak current with 3.3V voltage input. Power is always of concern for wireless medical telemetry system s A dual functio nal helix antenna is designed on a Rogers RO3003 substrate with a dielectric constant of 3.0 and a thickness of 10 mils. To form a helix shape the antenna is composed of a piece of microstrip feeding line with a width of 0.65 mm and a titled trace with a length of 220 mm and a width of 0.3 mm. A tilting angle of 1.9 is used for the desired helix with a pitch of 0.63 mm. This antenna operates at 150 kHz for near field wireless charging while at 400MHz it works as the wireless communication antenna. Further, based on the tun a bility of metamaterial particles, a wireless battery free smart mouth guard is developed. The system is fabricated on a high frequency laminate (RO 3010, Rogers Inc.), which is based on ceramic filled Polytetrafluoroethylene (PTFE ) with a dielectric constant of 10.2, a loss tangent of 0.002, and a thickness of 10 mils (0.25 mm). The SRR is designed to resonate at a 5.8 GHz band, the outer diameter of the SRR is 6 mm (0.31 ), and the SRR trace width is 0.3 mm. A linear size reducti on up to 38% is achieved comparing to a half wavelength resonator. This system is not only compact but also battery free and wireless.

PAGE 18

18 CHAPTER 1 INTRODUCTION 1.1 Motivation With the growth of the demand for better life quality, more advanced healthcare tec Organization (WHO) survey, there is a global shortage of healthcare personnel. This fact is especially true in developing countries such that in Africa, 57 countrie s are in a critical shortage of healthcare workers, a deficit of 2.4 million doctors and nurses [ 1]. In developed countries, cost Australia, the healthcare cost for the year of 2005 2006 was as high as 87 billion Australia Dollar (AUD) which corresponded to 9% of its Gross Domestic Product (GDP) [2]. Biomedical healthcare devices fall into two fundamental catalogues: diagnosis devices and treatment devices. The maj or purposes for a diagnosis device is to detect and record various vital signs of a patient, and data of interest include electrocardiogram (ECG) waveform, blood pressure, temperature, glucose level, respiratory rate, blood oxygen content and so on. On t he other hand, more and more electrical devices are now being employed for treatment purposes. short) which is used to restart a heart that has stopped beati ng. Other devices such as drug delivery systems, electro thermal therapy machine, cardio pace makers and so forth are also becoming popular to relax certain symptoms. With the advances of wireless technologies and high performance, low power circuitry t echnologies, more and more medical systems are made wireless. A wireless signal acquisition system is often known as a telemetry system. This wireless method for vital data acquisition is greatly preferable for patients because of convenience in data acqui sition and transmission, in

PAGE 19

19 situ analysis capability, and better aesthetic outlook. However in these days, most of those They do not take full advantage of the powerful wireless ne twork, by which remote and in home monitoring would be facilitated. For this end there is a strong demand to develop low cost, reliable and compact healthcare devices with optimal performance. Meantime, mobile phones are becoming very common all around th e world. According to International Telecommunication Union (ITU), by the year of 2011, there are 331,600,000 cellular subscriptions in the United States which means 105.91 subscriptions per 100 in habitants. This trend is also true for developing countri es for example, in China, there are 73.19 cellular subscriptions out of 100 habitants. Total mobile cellular subscriptions reached almost 6 billion by the end of 2011, corresponding to a global penetration of 86% [3]. Mobile Health Technologies which are technologies are now of great research interest [4]. A n as the system coordinator. Due to the deep penetration of cellular network, mHealth systems are believed to have transformative force to healthcare providing infrastru ctures. A major design challenge of a mHealth system lies in its compact size and reliable performance. This is especially true for the radio frequency (RF) subsystem. Also, significant care should be taken to reduce power consumption, providing desired ra dio performance. Meanwhile, the system is supposed to meet certain regulations to be used, for biomedical devices. Especially, Federal Communications Commission (FCC) and Food and Drug Administration (FDA) have a series of requirements for the magnitude of RF power and frequencies in healthcare usage [5 ][6 ]

PAGE 20

20 FCC regulations focus on the compatibility of the device with other systems as well as the amount of radio power absorbed by the human body. FCC also regulates the spectrum allowed to be used by medica l devices and services. For example, the Bluetooth devices are assigned a band of 2400MHz to 2483.5MHz for operation and Medical Implant Communication Service (MICS) uses a band from 40 2 MHz to 40 5 MHz. Recently, the spectrum of 413MHz to 457 MHz has been assigned for medical micro power networks [5]. In addition, FCC has electromagnetic compatibility (EMC) regulations to all electronic equipment sold in the United States. Basically, a device is said to be electromagnetically compatible if (1) The device do es not generate interference to other systems; (2) The device is not susceptible to external emissions; (3) The device does not interfere with itself. Usually radiation and conduction emission tests need to be performed. For medical devices, a special char acteristic called Specific Absorption Rate (SAR) needs to be measured. SAR is an integral of electromagnetic (EM) field strength within a certain volume and it is a representation of how much EM energy is absorbed by human tissue s Meanwhile FDA regulatio n focuses more on the potential health risk that might be brought about by the devices. The a ntenna is one of the most important and challenging part s in a wireless medical device. Antenna synthesis usually involves the design of the following parameters : center frequency, bandwidth, radiation pattern, impedance, radiation efficiency, gain, and directivity. EM wave propagation and transmission through antennas is characterized by the Friis' equation: ( 1 1) where and are the received and transmitted power, respectively; and are the antenna gains of the transmitting and receiving antennas, respectively; is the wavelength of the radiat ed EM waves ; R is the distance between the antennas. In a certain propagation medi um wavelength

PAGE 21

21 is inversely proportional to the frequency, and so at a higher frequency, gets smaller, and the ratio of gets smaller which means a higher frequency EM wav e is more lossy comparing to a lower frequency wave at a given distance R Also, low frequency EM is usually less oxygen absorption and less lossy in space. This is why major communication systems today are using relatively low frequency. However, this rai ses a design challenge for antenna engineers S ince for most of the antenna types, antenna dimension is proportional to its working wavelength, a lower operation frequency requires a larger antenna physical size which makes it difficult to implement in a c ompact system. Therefore, significant research efforts have been exerted for antenna size reduction and electrically small antenna implementation For antenna s we have ( 1 2) where is the antenna aperture, G is the antenna gain. Equation ( 1 2) shows the relationship between the antenna aperture size and the antenna gain. It is obvious that at a fixed wavelength a smaller aperture corresponds to a lower antenna gain. Another performance concern for small antennas includes low bandwidth and low efficiency. Bandwidth requirements come from the communication protocol. To achieve low power consumption, simple modul ation schemes are usually used such as 2 FSK (frequency shift keying). An advantage of the frequency modulation lies in that the envelope is constant, it is not sensitive to amplitude distortion, and power amplifier (PA) in a transmitter can be driven to i ts 1dB compression point where the power efficiency is much higher than that in its linear region, by which the battery usage of the system can be saved. However, the frequency modulation has lower symbol efficiency due to its very simple constellation and it requires a broad bandwidth to operate. Bandwidth requirement also comes from spectrum spreading techniques. Such as in a Bluetooth system, frequency hopping spectrum spreading (FHSS) is used where the center frequency hops in a certain pattern,

PAGE 22

22 which g ives better signal to noise ratio (SNR) but does need relatively broad bandwidth to operate (about 100MHz). Antenna radiation efficiency is defined as the ratio of power radiated to the power delivered to the antenna which is equivalent to the ratio betwee n the radiation resistance and loss resistance of the antenna. Small antennas have usually small radiation resistance, while they show huge input reactance, either capacitive or inductive depending on different radiation schemes. The low radiation resistan ce and high input reactance makes small antennas difficult to match with desired characteristic impedance and low in radiation efficiency. An important consideration in antennas is their radiation directivity. An antenna is expected to have certain direc tivity, with which EM waves are directed from and to the desired direction with efficiency. In some cases, the orientation of the biomedical sensor system such as a wireless capsule endoscope, is not predictable and the blind spot of observation is a major concern. Thus for an antenna in such a system, an omnidirectional radiation pattern is more preferred to a directional one. Often, i ntegration of such an antenna with other system components could be difficult since the interconnection between different boards by multilayer vias may introduce additional insertion loss, and electromagnetic interference ( EMI ) from antennas may introduce distortion to the waveform from the neighboring electronic circuits such as high speed digital bus lines Also, harmonics and switching noise from high speed digital circuitry and power conditioning systems may bring in unwanted interference to the antenna and other radio frequency (RF) subsystems. With the advance of digital circuitry technologies, the digital bus clock fre quency is now approaching the microwave range (MHz) and therefore the EMI problem is becoming more significant. With the development of a numerical method, commercial electromagnetic simulation tools are becoming popular in assisting microwave device desi gn and optimization. Popular

PAGE 23

23 software includes High Frequency Structure Simulator (HFSS, Ansys Inc.) which is based on the Finite Element Method (FEM); Ansoft designer (Ansys Inc.) which is based on the Method of Moments (MoM); EMPro from Agilent which all ows a user to choose from different simulation engines. These tools are utilized for the design of antennas taking into account the antenna parameters mentioned above. The tools are especially useful when the antenna system is located in a harsh environmen t such as in a human body or a high temperature location where the direct antenna experiments are not economic and practical. 1.2 Microwave I nteraction with H uman B ody Before further discussion regarding wireless biomedical telemetry systems, it is important to study the interaction between EM waves and human tissue s Radio frequency behavior of a material is usually characterized by its electrical permittivity (in F/m), magnetic permeability (in H /m) and conductivity (in S/m). Human body subjects to an appli ed EM field need to satisfy certain boundary conditions, and the total field in the human body includes the induced field as well as the applied field. Usually, electric field strength is represented by E (in V/m) and another associated quality is called e lectric field flux density or charge density which is usually represented by D (in C/m 2 ). If the material is lossless, isotropic and linear, E and D can be associated by: ( 1 3) wher the electrical permittivity of the material, it is the product of the relative permittivity (or dielectric constant) and the electrical permittivity of free space Susceptibility of the material can be found from It should be noticed that all those discussions in this work are referring to the lossless, isotropic and linear media unless others are specified, and it is true for biological tissues [ 7 ]. As for a lossy media, we need to consider a phenomenon called

PAGE 24

24 polarization which results from the alignment of the molecule dipolar moment with the applied field. If polarization is delayed with respect to the variation of the applied field, the material is said to be lossy. If the polarization is not proportional to the applied field strength, the material is said to be nonlinear. If the polarization vector does not have the same direction as that of the applied field, the material is said to be anisotropic. Lossy nature of a material can be expressed using a complex dielectric constant: Loss tangent is defined as which is a representation of how lossy a material is. The if we consider a complex dielectric constant, is following: ( 1 4) We can define the effective conductivity of a material as While charges give rise to the electric field, the magnetic field is generated around a current carrying conductor. law states that the closed form integral of th e magnetic field strength H in A/m around a current carrying conductor equals to the total current flowing in the conductor. The total current includes conduction and displacement current components. For magnet ic fields, we define the magnetic current density B in Wb/m 2 as followings: H ( 1 5) where is the relative magnetic permittivity and is the magnetic permittivity of free space. Complex permeability can be defined to express loss as we have done for the electric field. The human body environment is composed of components with very different properties. Gabriel et al. have measured those constitutive parameters of a human body from 10Hz to 20GHz with an open end coaxial probe method and it has been illustrated that each different organ such as grey matter in brain, lung, heart and so forth ha s a totally different frequency

PAGE 25

25 response [6]. Thus for non radiating devices such as a transmission line or a coupler where a high frequency wave is mainly propagating locally, it is essential to consider the dielectric properties of the specific part of t he human body where this device is supposed to operate. However, according to FCC recommendation, radiating devices such as an antenna can be treated as a point source in the human body. In far field communication scenarios, the human body can be treated a s a med i um with a uniform dielectric property [7]. While characterizing a human body centered communication channel, it is always important to investigate penetration in biological tissues where a electromagnetic skin effect takes place. When a human body is exposed into an EM radiation, the responding fields, currents, and charges mainly concentrate near the surface of the human body. The distance the wave travel s in a lossy medium to its value of is defined as the skin depth of the and mathematically it can be calculated as followings [7] : ( 1 6) For good conductors , skin depth can be approximately expressed as while for a good dielectric where The skin depth can be described as a function of the the varies greatly from organ to organ, even for the same tissue, and the corresponding values vary over a wide range as shown in Table 1 1 [7]. Thus we cannot use the simplified skin depth equation of a good dielectric or a good conductor for human tissue study in most of the cases, but we need to stick to the original form of Equation ( 1 6 ) for skin depth calculation.

PAGE 26

26 Due to the complex nature of the human body tissue, it is always desirable to use a human body model or human body phantom for wireless biomedical device design and characterization. We mainly adopt the h uman body model of the Ansys HFSS in this work. The HFSS human body model has four different models for male and female with different modeling accuracy, which are: 1) Male model, full accuracy; 2) Male model, 2mm accuracy; 3) Female model, 2mm accuracy; 4) Male model, 4mm accuracy. The male model with 4mm accuracy is shown in Figure 1 1 It should be noticed that in this model, detailed muscle and bone structures are omitted I 50 F/m and a conductivity of 0.5 S/m. However, important organs such as lung, heart, colon, large intestine, liver and bladder are considered. For better accur acy, 2mm models for male and female are available as shown in Figure 1 2 In the 2mm model, important muscles are modeled. It also considers the difference between a male body and a female one. Further, there is a head only model available at this a ccuracy level which is very useful for hand held devices design such as a cell phone. It is also good for designing a medical device operating near or over human head such as a wireless EEG recorder. The 2mm head phantom is shown in Figure 1 3 In addition there is also a male model with full accuracy available. It details the whole human body in the microwave aspect.

PAGE 27

27 The h uman body model can assist the design procedure of an antenna operating in a human body centered network. For example, when we need to design a dipole antenna, we want to know the performance variation with the presence of a human body. We may simply setup as shown in Figure 1 5 with a 60 mm long dipole placed 30 mm away from the human head phantom. After simulation, the first phenomeno n we can observe is the radiation pattern change. A dipole antenna should have a donut shaped omnidirectional pattern in free space. However, with the human body, it shows a relatively directional pattern as shown in Figure 1 6. It can be seen that the rad iation is more directional toward the opposite direction of the human head. Another change we can see would be a center frequency shift from 2.03GHz to 1.9GHz. This is because the human body introduces a high permittivity medium to the free space and the e quivalent electrical length of the dipole gets larger which corresponds to a lower resonance frequency. Other parameter chang es are also summarized in Table 1 2 It can be seen that since the human body works as a microwave energy absorber, part of the radi ated power is absorbed by the human body, and radiation efficiency drops from 97% to 83%. Other parameters are also affected as shown. It should be noticed that not only an antenna but also other devices can also be simulated by this human body model. In a ddition to the antenna parameters, this human body model can also be used to simulate the SAR. As shown in Figure 1 7, if the input power to the antenna is 1W, the simulated peak SAR is 14 W/Kg over 1 gram of tissue which will fail the FCC regulation of 1.6 W/Kg. However, if we feed in 0.1 W, the measured peak SAR drops to 1.3 W/Kg which stays below the regulation. This human body model is also good for physical simulation, such as thermal simulation.

PAGE 28

28 Frequency selection is also important for human body c entered communication systems. Commonly used license free bands include the 400 MHz, 900 MHz and 2.4 GH z band. Higher frequency bands are available but a higher frequency EM wave results in more path loss. For a medical system, the link budget is usually v ery limited, and so we prefer to choose from relatively low frequencies such as those three bands. As shown in Figure 1 8, the 400 MHz band has the minimum path loss among those three license free bands and wave propagation properties vary very little for different people at this particular band. Thus the 400 MHz band is usually preferred from the performance point of view. However, the 400 MHz wave has a free space wavelength of 0.75 meter which is quite large and makes it difficult to design an antenna s mall enough to be implanted or worn by people. 1.3 Flexible E lectronics Flexible electronics refer to electronic systems based on flexible substrates. By stating ain degree. The most significant advantage of a flexible system comparing to its rigid counter parts lies in two points: 1) Flexible electronics could fit a human body profile better resulting in optimized wearing comfort. Also, those flexible parts are us ually easier to be implanted. 2) Flexible electronics offer the possibility to develop new architecture due to the flexibility such as a conformal structure which is not easily achievable with rigid electronic devices. The newly emerging structures introdu ce new properties such as shielding effects and also give a new method for size miniaturization for medical devices. In addition, other reasons such as the biocompatibility of certain flexible materials and its light weight nature all contribute to fast de velopment in flexible medical electronics. Due to

PAGE 29

29 those features, the medical application of a flexible device can be dated back as early as the Nowadays, not only are digital circuit boards made flexible, but also display units and sensors are flexible. In this work, we mainly explore flexible radio frequency circuits and their application to a medical system. From the hardware point of view, a system is usually compos ed of (1) an electronic substrate; (2) electronic components in a lumped and distributed form; (3) system packaging, or, encapsulation. In order to have a flexible system, all those parts need to be flexible to a certain extent. Flexible electronics can b e first implemented on a silicon base then transferred on to the implementation cost of such a method is usually high and it only covers a very small surface area and thus this method is usually found in a very small portion of a system board [12,13]. Instead, electronic systems could be fabricated on a flexible substrate directly with the recently introduced fabrication techniques such as additive printing [14], printing of etch mask [15], and the application of electronic function brought about by local chemical reaction [16] and so forth. In this work, we focus on the direct fabrication technique which is more straightforward and more suitable for large scale ma nufacturing. Due to the flexible behavior, there are more stringent electrical, biomedical, thermal and mechanical requirements for the substrate material used in a flexible medical system: (1)Dielectric surface evenness: Flexible substrates are usually thin and dielectric thickness variation is very critical especially for high frequency circuits to maintain their desired characteristic impedance. An even substrate helps to ease the fabrication procedure.

PAGE 30

30 (2)Biocompatibility: This mainly refers to the ca psulation layer, which interfaces with the human body directly. It should be flexible and does not interfere with the normal function of the surrounding tissues and organs. (3)Robustness: Flexible systems need to be physically bent, stretched or wrapped. T he dielectric and conductive layers are supposed to be robust enough to endure those mechanical variations. (4)Thermal expansion: Certain therapeutic devices operate at a relatively high temperature such as a device for thermal therapy. Thermal expansion needs to be taken into consideration in the design phase. It is usually desirable to have materials with a moderate or small coefficient of thermal expansion (CTE). A CTE of several ppm/C is used. With those parameters in consideration, liquid crystallin e polymer (LCP) is used as the flexible substrate in this work. LCP is a family of polymers which are extremely inert to most chemicals while showing relative low dielectric loss and moderate permittivity over a broad frequency span. LCP is synthesized by dissolving a polymer into a solvent or by heating a polymer above its glass or melting transition point [17]. Generally, LCP possesses all the properties that are typical for low molar mass liquid crystals in addition to new properties which cannot be obta ined from traditional low molar mass liquid crystals [18]. Nowadays, LCP products are commercially available in the form of double clad copper laminates. Ultralam 3850 from Rogers Corp. is a n LCP based substrate which has stable microwave performance up t o 10GHz [19]. The near ly unvarying dielectric constant eases the design procedure for high frequency circuits such as transmission lines and antennas. Flexibility of such a material comes from its very thin thickness. Thicknesses from 1 mil (0.025mm) to 4 mil s (0.1mm) are available to choose. Once designed and implemented on a flexible substrate, microwave circuits are

PAGE 31

31 usually in a 3 dimensional shape, and thus, those flexible microwave systems are also referred to as non planar structures. Special design c aution should be taken for flexible circuits. Among all sorts of flexible microwave circuits, passive antennas are of great interest. Special radiation performance such as antenna pattern variation and more design degrees of freedom are obtained once desi gned on a flexible base. Those new features will be detailed in later chapters. 1.4 Metamaterial and N on planar 3D A rchitectures for B iomedical A pplications Metamaterial is a family of artificial homogeneous structures that demonstrate unusual property that are not readily available in the nature. In microwave society, metamaterial refers to sub wavelength structures that show negative and/or near zero permit tivity and/or permeability. To be qualified as a homogeneous material, the unit cell size of such a structure needs to be smaller than a quarter wave length at the frequency of interest [20]. In a homogeneous material, the sub wavelength structures behave as a real material. As we will see in later sections, since metamaterial with simultaneous negative permittivity and permeability has antiparallel phase and group velocities and yields negative refractive index, it is also known as left handed (LH) materi al which is to compare with the traditional right handed (RH) material. Metamaterial can be physically implemented in a resonator form or transmission line form. Since metamaterial is composed of sub wavelength units, the most significant application is to use those particles as a sub wavelength resonator for microwave size reduction. Meanwhile, the transmission line based metamaterial demonstrates a nonlinear phase response, which makes it a good candidate for size reduction and multiple band operation dev ices. Metamaterial unit cells are usually modeled as an LC resonance tank. An application together with tunable devices such as a varactor gives tunability to the structure, and thus many other applications are based on the tunability of the tunable metama terial structure. In this work, certain unique features offered by metamaterial particles are used for biomedical

PAGE 32

32 devices Also, s ub wavelength effects help reduce the device size while the tunability can be used for tunable device design as well as meta material based transducers and sensors for bio information acquisition. 1.5 Objective and T hesis S tructure This thesis focuses on the study of metamaterial particles and their application s into biomedical telemetry platforms, especially systems based on flexible electronics. This thesis is organized as shown in Figure 1 9. D iscussion begins with three key components of wireless biomedical telemetry: the newly proposed metamaterial particles RF subsystems including active and passive RF components, and the antennas. After the component level discussion, two system level case studies are discussed The first system focus es on a wireless endoscopic diagnosis system with high performance and lo w assembl y cost while the second one aims at a smart mouth guard system for the detection of Bruxism and its management The data obtained from the latter system is linked to a mobile system using a standard communication protocol and a n mHealth architectu re, i.e. health monitoring using mobile devices, is implemented for Bruxism and other sleep disorder in home monitoring and remote diagnosis. Finally the conclusion will be drawn.

PAGE 33

33 Table 1 1 Microwave characteristics of human body and head [7] Frequency (MHz) Head Head Body Body Permittivity Conductivity Permittivity Conductivity 150 52.3 0.76 61.9 0.80 300 45.3 0.87 58.2 0.92 450 13.5 0.87 56.7 0.94 835 41.5 0.90 55.2 0.97 900 41.5 0.97 55.0 1.05 915 41.5 0.98 55.0 1.06 1450 40.5 1.20 54.0 1.30 1610 40.3 1.29 53.8 1.40 1800 2000 40.0 1.40 53.3 1.52 2450 39.2 1.80 52.7 1.95 3000 38.5 2.40 52.0 2.73 5800 35.3 5.27 48.2 6.00 Permittivity is the relative permittivity, with respect to the permittivity of free space; conductivity in S/m. Table 1 2 Antenna parameter comparison: In free space and beside the human head model Parameter In free Space With Human Phantom Center Frequency 2.03GHz 1.9GHz Gain 2.15dBi 5.18d B i Efficiency 97% 83% Bandwidth 10% 14% Input Impedance 73+42.5j 2.71 5.37j

PAGE 34

34 Figure 1 1 4mm accuracy male human body model from Ansys

PAGE 35

35 A B Figure 1 2 2mm accuracy hum an body models from Ansys. A) Male. B ) Female Figure 1 3 2mm accuracy human body head model from Ansys

PAGE 36

36 Figure 1 4 Full accuracy human body model from Ansys Figure 1 5 Dipole antenna with 4mm human body head model

PAGE 37

37 A B Figure 1 6 Simulated 3D radiation pattern. A ) Dipole antenna in fr ee space. B ) Dipole antenna beside the human head model A B Figure 1 7 SAR simulation based on the huma n body mode l A) Input power = 1W. B ) Input power = 0.1W

PAGE 38

38 A B C Figure 1 8 Path loss at differen t frequency bands in human body. A) 400MHz. B) 900MHz. C ) 2.4GHz [9]

PAGE 39

39 Figure 1 9 Structure of this dissertation work Wireless biomedical telemetry systems based on flexible metamaterial circuits and advanced RF architectures Introduction Metamaterial Circuits RF Subsystems Antennas for Biomedical Applications Wireless Capsule Endoscope Smart Mouth Guard Conclusion Components Systems

PAGE 40

40 CHAPTER 2 METAMATERIAL CIRCUITS 2.1 Background and Theory As different kinds as shown in Figure 2 1. The four possible sign combinations w here first three kinds are readily available in nature, and the fourth one is not while such a History of investigating materials with negative constitutive parameters dates back to 19 68 when Russian physicist Viktor Veselago [21] investigated the possibility and potential applications of wave propagation in a media with both negative permittivity and permeability. The fir st physical implementation of metamaterial was demonstrated by Smith et al. in [22]. As shown in F igure 2 2 the concentric ring structures which are known as split ring resonators (SRR) are responsible for negative permeability ; the short pins give nega tive permittivity. Metamaterial structures are also implemented in transmission line form [20]. As shown in Figure 2 3 while traditional transmission lines are modeled as a series inductor and a shunt capacitor the metamaterial transmission line is model ed as a capacitor in series with an inductor in shunt. However, du e to the unavoidable parasitics pure left handed transmission lines are not easy to implement physically, and thus a composite right and left handed (CRLH) transmission line is widely accep ted as the transmission line version of metamaterial as shown in Figure 2 4 As shown in the figure, there are four lumped elements in a single CRLH transmission line unit cell. Left handed inductor (LL) and capacitor ( CL), right handed inductor (LR) and

PAGE 41

41 c apacitor (CR) [20]. At the low frequency range, left handed part dominants, the unit cell shows left hand feature s such as the relationship of the antiparallel group velocity and phase velocity H owever, since the left handed (LH) part has a configuration of high pass nature, it has a cutoff frequency, which is shown, as the LH gap in Figure 2 4 I n the same sense, the right handed (RH) section has a low pass nature and shows a band gap at higher frequency. In general, the series re sonance se and shunt resonance sh are different, so that a gap exists between the LH and the RH ranges. However, if these resonances are made equal, or are "balanced," this gap disappears, and infinite wavelength propagation is achieved at the transition 0 2.2 Split Ring Resonator (SRR) and C omplimentary Split Ring Resonator (CSRR) Research on ring resonator s is not a whole new topic A comprehensive work on ring circuits has been done by Chang [23]. However, people f ind that when it is made in a sub wavelength regime and a proper dent is applied, negative permeability can be obtained in an array of this structure [24]. A split ring resonator (SRR) is shown in Figure 2 5 SRR responses to the external magnetic field that is perpendicular to the SRR surface. Usually, it is desirable to have more than one concentric SRRs arranged together to enhance its response. It should be noticed that for simplicity purposes, SRRs are also seen in rectangular outline instead of circular rings. For circular and squ are single ring SRRs with the same linear dimension, metal characterizations and haps, the circular one shows a higher resonance frequency, which a single ring SRR is the basic implementat ion, most of the modern research es us e double ring SRR s instead as shown in Figure 2 6. The main reason is to enhance the coupling, especially in an array case And t his kind of double ring arrangement will offer a more robust peak [25].

PAGE 42

42 It is shown in [24 ] that when the excitation magnetic field is perpendicular to the SRR surface, this structure gives the plasmonic type permeability as a function of the frequency in the form of (2 1) where p is the pitch in the SRR array, a is the inner radius of the smaller ring, radial spacing between rings) is the magnetic resonance frequency and ( : metal resistance) is the damping factor due to metallic loss. Equation ( 2 1 r ] < 0. (2 2) where is known as the magnetic plasma frequency. It should be noticed that recent studies show that for the specific orientation of the SRR and the polarization of the incident EM wave, it is possible to obtain resonance to the electric field excitation [26]. This effect is (EEMR) The SRR is one of the most well known implementation s of metamaterial with a negative permeability while a thi n wire or rod is the basic structure wh ich can offer a negative equivalent permittivity. The SRR works as a L C resonance circuit in this case, and thus we may figure out that this kind of L C resonance circuit may take many other forms than the SRR. An alt ernative form of the SRR is known as the complementary split ring resonator (CSRR) which is a negative image of the SRR on ground plane as shown in Figure 2 7

PAGE 43

43 Different from the SRR which is a structure interact ing with the external magnetic field, the CSRR responses with the electric field applied. It is also shown in [27] that by placing the CSRR beneath a microstrip line loaded with proper gaps, a band pass filter response is obtained due to the left handed behavior. Other metamaterial structures su ch as chiral material s are also available [28]. This kind of metamaterial is very promising in the optical community due to its polarization effects on optical waves. To investigate the electromagnetic wave interaction with the SRR, following vectors need to be defined: external stimuli electric field intensity E, electric flux density D, electric polarization P. is the permittivity of the media, which is the product of the free space permittivity and the relative permittivity of the media. Assuming a homogeneous media, from Maxwell equation, it is known that: ( 2 3 ) It is shown that the equivalent permittivity of the media is determined as [29] : ( 2 4 ) So the eff ective relative permittivity is : ( 2 5 ) Here N is the number of particles per unit volume of the capacitor; is the atomic / molecular electric polarizability. Similarly, the equivalent permeability of a media can be found as [29] : ( 2 6 )

PAGE 44

44 where is is the atomic / molecular magnetic polarizability, H is the external magnetic field intensity applied. is the free space permeability and equals 2.3 Metamaterial Applications 2.3.1 As A Homogeneous M aterial As illustrated in Figure 2 8A, this double negative material is composed of SRRs which suppl y negative permeability as well as t hin wires which are responsible for negative permittivity. By combining those two negative parameters, a material (media) with a negative reflection index (n) is made. From Figure 2 8 B we can see the measurement setup B asically this is to demonstrate the well The EM propagation through the n ormal material follows the aw. The n ormal material such as Teflon shows a normal refraction angle as shown in Figure 2 8 C by the dashed line. However, with the metamaterial loading, this setup shows interesting phenomena as shown in Figure 2 8 C with the solid line i.e. a negative r efraction angle is observed. The only possible explan ation is that the refraction index, n, is also negative at this moment. With the advancing of the computing facility and technology, microwave and RF devices are usually designed using electromagnetic simulation software. Popular numerical techniques include Method of Moments (MoM), t ime domain methods such as Finite difference time domain (FDTD) methods, and finite element methods (FEM). A large number of commercially available simulati on packages are also available. Numerical study of electromagnetic waves interacting with the negative index materials using HFSS is reported in [30]. 2.3.2 Loading onto Hosting T ransm ission L ines The most straight forward SRR applications to the planar transmi ssion line can be found in [31]. SRR arrays are loaded beside a microstrip line as shown in Figure 2 10 [31] :

PAGE 45

45 Microstrip lines are widely used in the microwave planar circuit design and microwave integrated circuit (MIC) technology. Quasi transverse electr omagnetic (TEM) mode is dominant in the microstrip line. So that when SRRs are placed next to the microstrip line, the magnetic field will cut through the SRR s perpendicularly generating the equivalent negative permeability but only over a narrow band as s hown in Figure 2 11 If we take a closer look at this filter, the the resonance frequency, and thus the traveling wave along the microstip line turns into standing waves in resonators and a stop band is obtained. 2.3.3 Loaded into Hosting Bulky W aveguide The c onventional hollow metallic waveguide is one of the most important transmission line structures due to its high power handling capability and very high quality factor. It is interesting to investigate wha t will happen if the hollow waveguide is loaded with resonators. As to my knowledge, the first experiment of this kind was conducted by Marques et al [32] The waveguide used in [32] has a square cross section with an area of 6 mm by 6 mm and a cutoff freq uency of 25GHz. The SRRs inserted in have a diameter of 5.6 mm. A very significant phenomenon is shown in Figure 2 13 A transmission peak appears in the frequency range between 5.8 GHz and 6.5 GHz. It is really remarkable that a transmission band is obtained below the cutoff frequency. This experiment is repeated in a more comprehensive way in [33]. In [33], the SRRs are designed to resona te at 7.8 GHz and loaded into two different waveguides: For the first case, the SRR array is loaded into a wavegu ide with a cutoff frequency of 4.3 GHz, and thus the SRRs resonate at a frequency above the cutoff frequency and the second case is to repeat the experiment conducted in [32] i.e. the SRR array with 7.8 GHz resonance frequency is loaded in a waveguide with a cutoff frequency of 12.5 GHz.

PAGE 46

46 The resultant S parameters are shown in Figure 2 14 respectively. For the first case, a stop band occurs in the transmission spectrum while in the second case, a pass ba nd appears below the cutoff frequency and thus waveguide miniaturization is achieved. It should be noticed that the SRRs also have impact s on the cutoff frequencies. It has been shown in Figure 2 14 that the cutoff frequencies shift to the higher range fo r both cases. Considering a waveguide as shown in Figure 2 15, the Poynting vector travels along the z+ direction. In order to have the best response from the SRRs, the SRR array should stay along z+ direction. The permeability is a tensor: where equations, we have a group of linear first order differential equations: ( 2 7 ) ( 2 8 ) ( 2 9 ) Solving the wave equation with respect to the waveguide boundary condition, we can find the following allowable wave vector: [33] ( 2 10 ) Above the cutoff frequency, hence negative will result in negative and the evanescent mode wave propagation and a stop band are observed B elow the cutoff frequency, we only have the evanescent mode, thus and hence the negative will result in positive and a transmission band.

PAGE 47

47 Also an alternative explan ation is given in [32]. Marques et al. treat the SRR loaded waveguide as a dielectric material loaded waveguide H owever, the dielectric material has the negative permittivity and permeability. A cutoff waveguide could be regarded as h aving the effective permittivity of [32] : ( 2 11 ) ( 2 12 ) Equation ( 2 12 ) gives the effective plasma frequency of this waveguide loaded with material. The general form of the wave vector in the waveguide loaded with a homogenous material is ( 2 13 ) Substituting the effective permittivity in Equation ( 2 11 ) into Equation ( 2 13 ) : (2 14 ) This agrees with Equation ( 2 10 ). We may say those two methods converge with each other. The h ollow waveguide with a circular cross section loaded with metamaterial is reported in [34] in an infra red region. It has been demonstrated that simultaneously negative permittivity and permeability is observed. 2.3.4 SRR Based Tunable D evice A w aveguide loaded with SRR array s is also a good candidate for in situ tunable device s As shown in Figure 2 16, one of the arrays has a piece of magnetic material attached to its edge. Thus, the relative position of those two arrays can be changed by moving a piece of ferrite (or a permanent magnet) at the outside of the waveguide. Since we are working at

PAGE 48

48 microwave frequencies, the effects from magnetic materials can be ignored at such high frequencies. The large SRR has an outer diameter of 18 mm, while that of the small one is 10 mm. The width of traces and the gap of slots are 2 mm and 0.5 mm, respectively, for bo th SRRs. Each substrate has a relative electrical permittivity of 2.55 and a thickness of approximately 1.5 mm (60 mil s ). When the back to back stacked SRR arrays are inserted in the waveguide and the measured transmission coefficients are shown in Figure 2 17 When two arrays are in an aligned mode, an additional stop band shows between 5 GHz and 6.5 GHz with a center frequency of 5.8 GHz while it does not when they are in a total misaligned mode. The relative position shift is achieved by simply moving a piece of ferrite near the waveguide. T hose magnetic pieces have very few effects on the wave propagation in the frequency of interest. B y changing the relative position of two coupled SRR arrays, the mutual coupling factor is changed, so that a tunable band rejection filter can be implemented in a bulky waveguide. As the small SRR array is continuously shifted with respect to the large SRR array starting from the first alignment, another band gap shows when the next SRR units are aligned with each other since they are periodic structures. If we assign x 0 (= 25 mm) as the periodicity of the SRR array we can have a continuous transmission curve at 6 GHz as a function of the relative misalignment between the two arrays as shown in Figure 2 18. It can be see n from Figure 2 x 0 = 0 i.e. two SRR arrays are aligned, we have a minimum transmission at approximately 6 G Hz. With more misalignment, the interaction between the 0 = 0.5, the SRR arrays are totally misaligned and a minimum insertion loss is observed. After that, the large SRRs will interact with the next small SRRs, so that the forbidden band gap becomes deeper

PAGE 49

49 0 = 1, the SRRs are totally aligned with t he next counterpart of the other array, and thus another minimum transmission is observed. However, in this case, the SRRs at the end of the arrays do not have counterparts to interact with, so that x 0 = 0 (i.e. the first tota l alignment case) occurs. 2.3.5 SRR itself as A n Electrically Small and Tunable A ntenna The proposed SRR is excited by a monopole as shown in Figure 2 19. When the external magnetic field cuts the SRR plane, at a specific plasma frequency, the SRR will show the resonance behavior. The structure shows that t he outer diameter of the SRR is 12 mm the inner diameter is 10 mm, the SRR width is 0.5 mm, and the slots on both SRR inner and outer rings are all 1 mm wide. It should be noticed that the outer ring has t wo slots, where one exists inherently for a conventional SRR and the other is newly added for the varactor diode connection in this research. The structure is implemented on a radio frequency substrate with a dielectric constant of 2.33 and a thickness of 30 mils. The microstrip line is 2.2 mm in width to match with the 50 ohm characteristic impedance. A monopole is made to excite the SRR antenna at the end of the microstrip line which has a length of 13.3 mm and is supposed to radiate at 4 GHz. To enhance the antenna performance at 2.1 GHz, a section of a ohm microstrip feeding line and the monopole. This SRR antenna is first simulated using the HFSS. Figure 2 20 shows the reflection coefficient (S11) with differen t varactor values between 0.76 pF and 4.15 pF. The plot shows two discrete resonance bands: the lower one due to the radiation from the SRR antenna varying from 2.1 GHz to 1.7 GHz with a percentage change of approximately 20% for different varactor values and the other one around 3.8GHz due to the exciting monopole, relatively little changing from 3.8 GHz to 3.6 GHz with a percentage change

PAGE 50

50 of approximately 5%, attributable to the impedance changing of the whole radiating system at different capacitance val ues. A tunable SRR antenna is investigated both theoretically and experimentally. The proposed antenna has a variable radiation frequency from 1.81 GHz to 2.14 GHz with a tuning voltage between 0 and 9.5 V. Simulated data show good agreement with measureme nt ones. Besides tuability, the proposed antenna has achieved a linear dimension reduction by 59.3 % compared with the monopole counterpart. 2.3.6 Other A pplications SRR and CSRR particles are also used to improve antenna performance, implement RF and optical cloaking, act as absorbing material up to Terahertz range and so forth. We will revisit the CSRR based antenna size miniaturization in later chapters.

PAGE 51

51 Figure 2 1 Materials of different kinds [20] Figure 2 2 Metamaterial with simultaneous negative permittivity and permeability Figure 2 3 Left handed transmission line [20]

PAGE 52

52 A B Figure 2 4 CRLH model. A ) Composite right an d left handed transmission line. B) CRLH dispersion relation [20] Figure 2 5 Split r ing r esonator (SRR) and its equivalent circuits Figure 2 6 Double SRR and its equivalent circuits

PAGE 53

53 Figure 2 7 Complementary s plit r ing r esonator (CSRR) and its equivalent circuits [27] A B C Figure 2 8 Negative refraction index. A )Metamaterial wi th negative refraction index. B) Measurement setup. C )Measure d transition power [30]

PAGE 54

54 Figure 2 9 Simulated wave propagation through a material with negative refraction index [30] Figure 2 10 Microstrip Loaded with SRR Arrays Figure 2 11 S parameter of SRR Loaded Microstrip [31]

PAGE 55

55 Figure 2 12 Hollow Waveguide Loaded with SRR Array [32] Figure 2 13 Transmission Coefficients (S21) of Waveguide Loaded with SRR Array Figure 2 14 Waveguide Loaded with SRR

PAGE 56

56 Figure 2 15 Waveguide c onfiguration A B C Figure 2 16 Waveguide loa ded with SRR for in situ tuning. A) Simulated aligned mode. B) Simulated Misaligned mode. C ) Fabricated SRR arrays x y z a

PAGE 57

57 Figure 2 17 Measured in situ t u nability of the proposed waveguide Figure 2 18 Tra nsmission at 6 GHz with respect to different misalignment level A B Figure 2 19 Proposed electrically small and tunable antenna A) Schematic. B ) Fabricated Device -100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 2 3 4 5 6 7 No Filling Aligned Mis-Aligned Frequency/GHz S21/dB -50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0 0 0.2 0.4 0.6 0.8 1 S21 at 6GHz x/x 0 PEC Monopole Varactor

PAGE 58

58 Figure 2 20 Simulated return loss of the SRR Antenna

PAGE 59

59 CHAPTER 3 RF SUBSYSTEMS 3.1 Microwave Board Level Design Considerations Gigahertz circuit board design is one of the major challeng es for a wireless biomedical platform. The major difficulties come from the high performance requirements and limited available resource ( small board area, limited available power, complicated operation environment and so forth). Meanwhile, a commercially su ccessful product needs to be certified by related authorities such as FCC and FDA as stated in the previous chapter. Comparing to a low frequency or DC board, the most significant difference of the high frequency board is that the trace width and board th ickness need to be chosen in a certain manner so that the reflection loss is minimized while the high frequency communication subsystem needs to be isolated from other parts such as digital bus and switching power regulation circuitry in order to avoid int erferences. Finally, the radio frequency (RF) board is usually shielded to pass the FCC regulations regarding radiation emissions and to further isolate the high frequency part from other sections. An RF designer needs to determine which kind of transmissi on line technology is going to be employed in the system. The main function of the transmission line is to deliver a desired amount of power from the source to the load in a proper way. For low frequency circuits where the wavelength of the signal is typic ally much greater than the physical length of the transmission line, one can simply use wires to connect the components on a circuit in a desired way. However, as frequency goes higher, the wavelength of the signal is comparable to the transmission line le ngth, and significant reflection power loss and dispersive effects take place. Dispersion on a transmission line introduces waveform distortion to the signal that is traveling along the line.

PAGE 60

60 To establish desired communication between different nodes on a circuit board, several different types of transmission lines have been introduced as shown in Figure 3 1 [35] Among all those physical forms, micr ostrip lines and co planar wave guide s (CPW) are the most popular ones T he main reason is that those kinds of transmission line structures are compatible with the standard print circuit board (PCB) fabrication procedure and can be used for mass production. Although the transmission line takes different physical forms, it is usually represented by two parallel lines in schematic regardless of the actual shape of the line. The lumped element model of a transmission line can be represented as Figure 3 2 It should be noticed that for the different implementation of transmission types, different line parameters a re expected. Usually, those line parameters are associated with (1) the geometric configuration of the transmission line and (2) the constitutive parameters of the material of the transmission lines. So the transmission line design procedure is to choose a proper material and a n appropriate corresponding geometric profile. We assume that the signal is traveling along the z direction on a transmission line I f we Figure 3 2, and let v (z,t) and i (z,t) r epresent the voltage and current distribution on the line respectively, by applying the l a w, we have : ( 3 1) I f we choose we have ( 3 2) I f we take phasor representation, that is, if we let ( 3 3)

PAGE 61

61 T hen ( 3 4) S imilarly, for current, we have: ( 3 5) S ubstitut ing ( 3 4 ) into ( 3 5 ) we can obtain the wave equation for : ( 3 6) A s imilar equation holds true for current, we define propagation constant as: ( 3 7) A ttenuation constant is a loss associated parameter in Np/m while is called the phase constant and its unit is rad/m. The wave equation can be expressed in the following form: ( 3 8) ( 3 9) where r epresents the wave propagating in the +z direction and represents the wave propag ating in z direction. From Eq uation ( 3 4) we can associate and as: ( 3 10 ) where we can define the characteristic impedance of a transmission line Z 0 as: ( 3 1 1) It can be seen that the characteristic impedance stands for the ratio between the amplitude of the incident or reflec ted voltage and the corresponding current. If we consider the lossless case which is approximately true for many of the cases, we have for the TEM

PAGE 62

62 mode or quasi TEM mode transmission lines, and we further have while the lossless character istic impedance is: (3 1 2) The wavelength for a material with the and the unit permittivity ( = 0 r ) : (3 1 3) is the free space wavelength. It is obvious that the effective wavelength gets smaller if a material with a larger dielectric constant is used as the propagation media. This leads to a popular way to reduce t he physical size of a microwave device. To achieve the same electrical length, the physical dimension of an element in a high dielectric constant material is much smaller than that of the same device based on a low dielectric constant substrate. However, h igh dielectric constant material s usually demonstrate a higher dielectric lossless assumption invalid Also a high dielectric constant substrate has a very limited tolerance to the fabrication error, and thus an accurate fabrication facility is a must to process devices on high dielectric constant material s (such as ceramic). If transmission loss is ignored, the most significant parameter preventing the wave from propagating properly is the reflection. Reflection on a tra nsmission line can be characterized by a factor the so the reflected voltage and the amplitude of the incident voltage. We assume the transmission line is terminated by a load with an impedance of Z L and the reflection coefficient is defined as: (3 1 4)

PAGE 63

63 It should be noticed that it is possible for both and to be complex numbers. And a load is said to be a matched one if in which case no reflection takes place on the load side. The voltage maximum on a standing wave pattern is caused by the phase add up from the incident and reflected wave components while the voltage v a lley comes from the out of phase canceling out of the voltage components from both waves. It should be noticed that the repetition period of the standing wave is half of a wave length. Standing wave can be characterized by the voltage standing wave ratio (VSWR) which is d efined to be the ratio between the amplitudes of the voltage maxima and minima The VSWR can be associated with the reflection coefficients in the following way: (3 1 5) Since the VSWR is a ratio, it is a dimensionless parameter. After proper geometric configuration and material is chosen, one needs to place the RF subsystem smartly on the board to achieve best overall performance. The major concern is the EMI problem, as stated in Chapter 1 T he system is not supposed to interference with neither itself nor other systems placed nearby. Usually, high speed digital buses should be placed far away from RF circuits, and this is mainly because nowadays, the digital bus speed i s approaching the microwave range (MHz or GHz), which will introduce interference onto the radio frequency structure. Even for a low speed digital bus, its harmonics could be a concern for RF designers as well. For antenna s it is more obvious because that interference from surrounding circuits could be considered as spurious feeding, which may excite cross polarized mode s Meanwhile, those circuitry around an antenna may work as a microwave absorbing material which will take away part of the radiated power thus degrade the antenna performance,

PAGE 64

64 such as a shifted input impedance and a very low radiation efficiency. On another hand, the radiation from microwave circuits would be coupled to digital buses, causing waveform distortion and the increase of the bit error rate (BER) ultimately T his phenomenon will bring Power conditioning circuits such as DC/DC converters are now a must in medical circuits T hey guarantee a stable power supply to high precision analog and digital circuits, and some circuits can extend the battery life by up converting the battery output voltage. Since it is difficult to effectively do voltage regulation with the DC signal, most of those power conditionin g systems are based on high speed switching circuits where switching noise is not avoidable. Layout of those high speed and high power subsystems could be challenging as well P roper isolation must be taken in order to prevent the switching noise from be i ng coupled onto the microwave circuits. With a limited board area, it is common to use via for inter layer communication. However, high frequency via should be avoided whenever possible T his is because a via structure is more susceptible to external interference S ince the characteristic impedance of a via is not easy to determine, the via structure usually introduce s additional insertion loss due to impedance mismatching. Further, at high frequenc ies via structures offer complicated parasitics and a re difficult to analyze and design. For certain transmission technologies such as for a coplanar wave guide with a ground plane, a considerable amount of vias need to be placed between the top side ground and the back ground T he main purpose to do so is to suppress the higher modes and ground the interference signal s from other parts. Usually, this is a good solution for the frequency signals with many

PAGE 65

65 ripples As a general rule, an immediate via is suggested wherever a ground is needed. This is to minim ize the parasitic components on a via pad. If a turning is needed when routing high frequency traces, it is always necessary to avoid 90 degree corners which would work as a good antenna for noise picking up, while those sharp corners resonate strongly at a certain frequency. Us ing a 45 degree corner a round corner or a truncated corner whenever possible is recommended as shown in Figure 3 3 [35] When routing a DC power supply to an IC or other components, it is necessary to place decouping capacitors. O ne of the functions of those decapouling capacitor s is to offer a relative ly stable voltage when a current draw on a certain pin changes rapidly. This happens in a certain scenario such as a microprocessor wakes up from a sleep mode when the current draw rises suddenly while usually the voltage supply needs certain time to respond this change A decoupling capacitor may assist to maintain a relative ly stable voltage in such cases. Another function of these capacitors is to decouple the co upling between different systems. Power traces may pick up microwave signals and those coupled signal s may enter the DC source which may cause unwanted short circuit s on the DC power supply. The d ecoupling capacitor offers a low impedance path to ground fo r those unwanted couplings and gives a pure path to DC power. It is thus reasonable to assume that RF circuits should be kept far away from power supplying pins, and decoupling capacitor s should be used as much as possible and should be placed as close to the power supply pin s on the IC as possible. In addition, power traces should be made as wide as possible to give a low impedance path for the DC power and wider trace s will not constrain the current level flowing on the trace. In addition to all those ele ctrical consideration, certain thermal and mechanical design rules are to be followed for the high frequency layout. In most design s a thermal pad is placed

PAGE 66

66 especially for high power scenarios such as for the power amplifiers ( PA ) in a RF transmitter. In some designs, a large thermal pad works as both a ground plane as well as a thermal relief component I n such a case, more than one via should be used on the pad for reliable system performance. If the board is subject to a thermal reflow procedure, prope r space should be designed between components to allow enough heat flow according to the desired temperature profile while soldering. As for different components, different transmission line technologies might be used in a hybrid way I n such cases one m ay need to design a transition between different feeding n systems, due to the availability of high performance surface mount (SMT) type components, it is popular to design Baluns with lumped SMT parts I n that way, it is usually desirable to place all parts as close to each other as possible which is to minimize the parasitic s from the interconnections and thus reduce the insertion loss and unwanted frequency response. 3.2 RF Transceiver Modern RF transceivers usually take the form of super heterodyne architecture. As shown in Figure 3 4 In a superheterodyne receiver, the receive d signal will be amplified by a front end amplifier W e will have a detail discussion on the amplifier in later sections of this chapter. The amplified signal is then mixed within a mixer with a locally generated demodulation signal. Most of mixers are bas ed on a nonlinear Schottky diode whose input and output characteristics can be expressed as a Tayler series: ( 3 16 )

PAGE 67

67 where is the output term and are input s to the amplifier, and we assume that and and then the second term can be expanded as (3 17 ) (3 18 ) This is the term that gives the sum and differential terms of the input frequencies. So that the output spectrum looks like Figure 3 5 If we revisit the Tayler series, there are still many higher order terms left, and those higher order terms will generate hi gher order harmonics which a re usually considered as distortion in a communication system. In a demodulator, we may be only interested in the frequency term of which is known as the intermediate frequency (IF) while another frequency sum term is known as the image frequency which needs to be filtered out by a band pass filter. After the band pass filter, we are only supposed to have the IF signal left in the system, and an IF amplifier is then used to amplify the IF signal to a proper level for d emodulation. Demodulation could also be done based on a mixer. However, for phase modulations, I and Q channels need to be demodulated separately and the balance between two channels is very important. After demodulat ion the signal frequency is low enough to be processed directly by a digital circuit, and this signal is known as the base band signal. A m ixer is usually characterized by the conversion loss, the 1dB compression point, the isolation and noise figure. We define the power level difference betwe en the input RF power and the output IF power as the conversion loss of a mixer I f we fix the local oscillator output power at a certain value and sweep the input RF signal power, we can see the output IF power saturate at a certain level W hen the IF power de viates from its linear region by 1dB, the 1dB

PAGE 68

68 compression point of a mixer is reached, below which we have the mixer working in a linear region. Another parameter which is important but sometimes ignored is the isolation between the local oscillati on port and the input RF port. This is of practical importance for the system to pass FCC EMI radiation emission. If the isolation is poor, the local oscillation signal might be coupled to the RF path and generate unwanted emission. Finally, we need to cha racterize the noise contribution of a mixer A lthough the noise figure of a mixer is not as important as that of a front end low noise amplifier, it contributes to the overall system performance as well. Conversion loss and 1dB compression point can be ch aracterized by using signal generators and spectrum analyzers The noise figure measurement does require a noise source and a noise figure analyzer. 3.3 Amplifier In a wireless communication system, the received signal is usually noisy and weak. The first bui lding block of a wireless receiver is an amplifier as shown in Figure 3 4. In a cascaded system, the noise performance is mainly determined by the first stage S ince this amplifier stays at the very front, its noise performance is critical to the overall system sensitivity. The major task for an amplifier is to effectively amplify the signal amplitude with minimum distortion on the waveform. An a mplifier shows nonlinearity at a certain input power level, which is similar to a mixer. The nonlinearity of an amplifier can be characterized by its 1dB compression point as well. The input/output characteristic curve of an amplifier is shown in Figure 3 6 As we have defined for a mixer, 1dB compression point is found where the actual output power derives from its linear region by 1dB. Another important factor for the amplifier characterization is called the 3 rd order intermodulation point (IP3). This parame ter determines the intermodulation behavior of an amplifier. When two signals with different frequencies are fed into an amplifier simultaneously, intermodulation takes places. The the 3 rd modulation

PAGE 69

69 product usually refers to the signal component with the frequency of Excessive intermodulation products will degrade the system performance. It is important to wisely choose the operation point for an amplifier. The o peration point close to the 1dB compression point offers greater power efficiency S i nce the gain is at its maximum but the signal amplitude compression is observed certain waveform distortion is expected However, for frequency modulation such as GMSK (Gaussian Minimum S hift Keying) in GSM ( Global System for Mobile Communications ) system the amplified distortion does not bother the system performance since the signal envelope is constant. For phase modulation such as QPSK (Quadrature Phase Shift Keying) where signal amplitude also carries the modulation information, it is thus undesirable to operate at the nonlinear region. In phase modulation systems such as in EDGE (E nhanced D ata R ate for GSM E volution ) it is always necessary to operate the amplifier in the linear region. In a broadband system such as a spectrum analyzer, another desig n parameter needs to be kept in mind is the gain flatness. A b roadband amplifier often shows a gain roll off at the higher frequency T his is mainly due to the packaging parasitics. A flat gain can be implemented by an optimized matching network [35]. However, this method requires a complicated design procedure and therefore design complexity is increased [36]. Feedback is also a solution to the gain roll off [37], but the feedback network needs extremely careful tuning and the bandwidth is usually limited and may not be suitable for broadband preamplifier applications. Here we focus on the preamplifier with a frequency spectrum of 100MHz to 7GHz, which covers many interesting industrial, scientific and medical (ISM) radio bands such as 433 MHz (Industrial usage), 2.45GHz (Wi Fi, Bluetooth) and 5.8GHz (Wi Fi, cordless phones), and

PAGE 70

70 402 MHz (Medical Implant Commission Service, MICS band) etc. Also, this covers frequency bands for GSM or CDMA (Code Division Multiple Access) cellular services. In this work, a compact passive frequency equalizer which is easy to design and cost effective to implement is presented. The equalizer is designed to wo rk over a wide frequency range of 100MHz to 7GHz. The targeted gain variation within this range is as small as 1.1dB. The proposed equalizer is based on a T attenuator. With a proper modification, it can compensate the amplifier gain roll off over the fre quencies. As a testing vehicle for this equalizer, a three stage amplifier consisting of three NBB 400 based gain blocks (RFMD Inc.) is implemented and cascaded with the designed equalizer. Amplifier performance with and without the equalizer is compared. The targeted technical specifications of a broadband amplifier in this work are: (1) Operation bandwidth: 100 MHz to 7 GHz; (2) Nominal gain: greater than 30 dB; (3) Gain variation: 1.5 dB; (4) Reverse isolation: greater than 40 dB; (5) DC current draw: less than 180 mA at +5 V; Taking all those points into consideration, an RF transistor, NBB 400 from RFMD is chosen as the main gain block. A single NBB 400 gain block offers a nominal gain of 15.5 dB with a 3 dB bandwidth of 7.5 GHz. For normal ope ration, this gain block draws a current of 70 mA with a noise figure of 4.3 dB at 3 GHz. A preliminary test with the single stage gain block shows a gain roll off of 4 dB from DC to 7GHz, the single stage amplifier is shown in Figure 3 7

PAGE 71

71 The single stage amplifier is then characterized with an Agilent E5071C vector network analyzer (VNA) after a standard two port calibration from 100MHz to 7GHz. The measured gain is shown in Figure 3 8 It is clearly seen that the single stage amplifier shows a gain roll off from 16.3 dB at 105 MHz to 10.9 dB at 7 GHz resulting in 5.4 dB gain variation. To implement a high gain amplifier, multiple stages of the NBB 400 based amplifiers are cascaded. The NBB 400 has high input impedance, and it is directly cascaded without loading effect. In this work, three stages of the NBB 400 based amplifiers are implemented as shown in Figure 3 9. A list of components used for this amplifier is shown in Table.3 1. The fabricated device is characterized using a VNA (Agilent E5071C) after a standard two port calibration between 100 MHz and 7 GHz. The measured gain is shown in Figure 3 10. Compared with the single stage case in Figure 3 8, the gain roll off of the 3 stage one is more significant. The peak gain of 46.97 dB is achieved at 400 MHz while the valley gain is only 34.7 dB at 6.8 GHz, resulting in an overall gain variation of 12.27 dB. At a DC bias of +5V, the total current drawn from the power source by the 3 sta ge cascaded amplifier is 157 mA which meets the targeted design specification. In order to compensate the gain roll off, increas ed attenuation at high frequencies is desired. For this, a shunt inductor and a parallel capacitor are added onto a T attenuator as shown in Figure 3 11 where and [35] where Z 0 is the characteristic impedance of the system in o hms, A is the attenuation level of interest in decibel. Since ideal resistors are not frequency dependent components, the attenuation is assumed to be invariant over the operation frequency span.

PAGE 72

72 At a low frequency, the impedance of the bridging capacitor C 1 is very large and the circuit behaves as the conventional T attenuator while at a high frequency its impedance beco mes very small and all the RF signals bypass the T attenuator, resulting in a reduced attenuation. Meantime, L 1 works in an opposite way I t offers a low impedance path for a low frequency signal to the ground while it shows very high impedance at a higher frequency. With appropriate values of C 1 and L 1 nonlinearly increasing attenuation can be achieved as the frequency increases For example, if a 10 dB T attenuator is selected, R 1 = 26 o hms and R 2 = 35 o hms. Further C 1 = 0.5 pF and L 1 = 1 nH produce a r ampin g attenuation response as shown in Figure 3 12. With simple modification, a normal T attenuato r shows attenuation ramping in the frequency domain. The curve shown in Figure 3 12 has an opposite trend to that of the curve in Figure 3 10. The principle is to connect the modified T attenuator with the cascaded multiple stage amplifier in series to compensate the gain roll off and ultimately achieve good gain flatness over the overall frequency range of interest The design of a frequency equalizer is performed using HFSS. The design procedure of a frequency equalizer begins with the measurement of the amplifier gain over the desired frequency range. The difference between the desired gain and the measured gain is calculated, where this ga in difference will be an off set to be compensated using the modified T attenuator, i.e. the passive frequency equalizer. With the desired frequency response as a reference, careful tuning of the passive dimensions is performed via HFSS simulation to obtai n a desired response. From the gain response of the 3 stage amplifier in Figure 3 10, where the minimum gain is 34.7 dB over the frequency range. So, a gain of 33.5 dB 1.1dB is set as a targeted gain. The

PAGE 73

73 difference between the desired gain and the measur ed gain in the frequency range of 100 MHz to 7 GHz is shown in Figure 3 13. Since we are going to use a T attenuator with a ramping frequency response described in the previous section, we need to choose appropriate values of the capacitor, inductor, and r esistors. Precision (1%) 0402 packaging surface mount type resistors are used to implement the attenuator. Capacitance and inductance can be implemented by either surface mount type lumped elements or distributed microstrip lines. In this work, a surface m ount type capacitor with 0402 packaging and a distributed microstrip line inductor have been used. The response at the very beginning of the frequency equalizer (for example at 100MHz in this case) is solely determined by the attenuation of the T attenuato r. With a desired attenuation of 14.7 dB, the calculated values of R 1 and R 2 are 18 o hms and 11.3 o hms, respectively. Low frequency response is also affected by L 1 To implement this inductor, a piece of microstrip line with a dimension of 0.3mm0.5mm is p laced between R 2 and the ground. As for reduced attenuation in high frequency, a C 1 value of 0.4 pF has been chosen for the best gain flatness over the frequency. The simulated frequency equalizer response is compared with the desired response in Figure 3 14. It can be seen from Figure 3 14 that with the proposed values, the simulated frequency equalizer shows good agreement with the desired response curve obtained from the amplifier measurement. The frequency equalizer is then fabricated on a same substrat e, RT/Duroid 5880. The frequency equalizer is then cascaded with the implemented 3 stage amplifier shown in Figure 3 15. It can be seen that after cascading the amplifier with the frequency equalizer, the maximum gain occurs at 5.3 GHz, with a gain of 34. 6 dB while the minimum gain of 32.4 dB is

PAGE 74

74 observed at 455 MHz. The gain variation is as small as 1.1dB over the frequency range from 100MHz to 7GHz. However, the tradeoff is to sacrifice certain amplifier gain for response flatness. The maximum gain varia tion occurs between 4 GHz and 6 GHz range, which is consistent with the observation in Figure 3 16. This variation is attributed in part to the tolerance of lumped element specifications and fabrication imperfection when a milling machine is used for patte rning. Using a more accurate fabrication method such as UV lithography patterning, with which a better gain flatness is expected, can relieve the latter Here, a compact, low cost frequency equalizer is designed fabricated and characterized, and the design guidelines are given. With an area of 6 mm 8 mm, a gain flatness of 1.1 dB is demonstrated from 100MHz to 7GHz with a nominal gain of 33.5dB. The proposed frequency equalizer is very easy to design and implement.

PAGE 75

75 Table 3 1 Bill of material (BOM) of the cascaded ampli fier Part Reference Description Manufacture Manufacture Model C 1 C 8 C 11 Capacitor Tantalum 4.7 F AVX Corp. TAJC475K025RNJ C 2 C 7 C 10 Capacitor 1000 pF TKD Corp. C1608X7R1H102K C 3 ,C 4 C 6 C 8 Capacitor 1 F TKD Corp. C1005X5R0J105K R 1 R 2 R 3 Resistor 22 o hm s Panasonic ERJ P14J220U L 1 L 2 ,L 3 RF Choke Minicircuits ADCH 80A Q 1 Q 2 Q 3 Gain block RFMD NBB 400 Substrate 31 mil Rogers RT/ D uroid 5880 A B C D Figure 3 1 Different transmission lines A) Microstrip. B) Strip line. C) Coplanar waveguide. D ) Slot line Figure 3 2 Lumped element model of a transmission line[35] Figure 3 3 Different microstrip bend z

PAGE 76

76 Figure 3 4 Superheterodyne Receiver Figure 3 5 Intermodulation Figure 3 6 Amplifier power sweep response [35]

PAGE 77

77 Figure 3 7 Single stage amplifier and schematic Figure 3 8 Measured single stage amplifier performance 0 4 8 12 16 20 0 1 2 3 4 5 6 7 Frequency / GHz Gain / dB ADCH 80A NBB 400 Capacitors Resistor

PAGE 78

78 A B Figure 3 9 Cascaded amplifier. A) Schematic. B ) Fabricated device Figure 3 10 Frequency response of the cascaded amplifier 30 32 34 36 38 40 42 44 46 48 0 1 2 3 4 5 6 7 Frequency / GHz Gain / dB

PAGE 79

79 Figure 3 11 Frequency leveling network s chematic and layout Figure 3 12 Frequency leveling network performance Figure 3 13 Desired frequency response of the frequency equalizer -12 -10 -8 -6 -4 -2 0 0 1 2 3 4 5 6 7 Frequency / GHz -16 -12 -8 -4 0 0 1 2 3 4 5 6 7 Frequency / GHz Transmission Coefficients / dB

PAGE 80

80 Figure 3 14 Comparison of the desired response and simulated response of the frequency equalizer Figure 3 15 Fabricated broadband amplifier with the frequency equalizer Figure 3 16 Measured amplifier gain with the f requency equalizer -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0 0 1 2 3 4 5 6 7 Desired Simulated Frequency / GHz Transmission Coefficients / dB 29 30 31 32 33 34 35 36 0 1 2 3 4 5 6 7 Frequency / GHz Gain / dB

PAGE 81

81 CHAPTER 4 ANTENNAS FOR BIOMEDICAL APPLICATIONS 4.1 Micristrip Patch Antenna : Design and Applications With the advances in biomedical electronics technologies, more and more wireless biomedical devices are emerging. The a ntenna used in such a system plays a similar role as it do es in other wireless communication systems: to delivery as much as energy from source to destination. However, as stated in Chapter 2 the human body offers a very complicated communication channel, and the antenna design for biomedical application is much more challenging. Depending on different applications, the antenna for medical therapy can be classified as two major categories: (1) Antennas for therapeutic applications and (2) antenna s for dia gnostic imaging and sensing. For the first case, localized heat is applied to the desired portion of the human body for treatment purposes while in the latter case antenna s are usually integrated with certain sensing or imaging systems to form a complete c ommunication system for patient health monitoring and disease management. The p atch antenna is good for both applications. The p atch antenna is one of the most wide ly used antenna forms. Its popularity is mainly due to its easy design procedure, low imple mentation cost, light weight and so forth. Major operation disadvantages of a patch are their low efficiency low power handling capability, high Q sometimes in excess of 100, poor polarization purity poor scan performance spurious feed radiation and v e ry narrow frequency bandwidth [38] A patch antenna is shown in Figure 4 1 A patch can be modeled as an open ended resonance tank, the resonance frequency ( 4 1 ) For Microstrip antennas, the dominant mode is TM010 with the next high order mode is TM001, For the dominant TM010 mode, the resonance frequency

PAGE 82

82 of a microstrip antenna is usually a function of its length. The p atch length is usually longer tha n the half wavelength by a fringing factor. The f ield distribution on a patch is demonstrated in Figure 4 2 To design a patch antenna, we need to first specify the center frequency f dielectric h T hen the patch length L can be determined by Eq uation ( 4 1 ), and the patch width is chosen to have a desired input impedance. Input impedance matching can be further improved by varying feeding scheme s or the position s of the feeding point. 4.2 Wrapped Patch An tenna with Omni directional Radiation Pattern The a ntenna s for biomedical implants are challenging to design, and the major concerns include: 1) Limited board space require s small antenna footprint; 2) Stringent SAR requirements need careful EMI design; 3) Limited power supply requires the system has an antenna with good impedance matching and radiation efficiency; 4) I n some cases such as a capsule endoscope application the orientation of the sensor is not predictable, and it is desirable to have an omnid irectional radiation pattern; 5) Sensitive digital circuits and sensors should be well isolated from interference of the antenna and other system in the environment; 6) It should be easy and cost effective to integrate the antenna with the system while maintaining reasonable performance It is not easy to achieve all those stringent requirements all in the same time especially for the microstrip patch antenna. A conventional patch antenna is constructed by two metallic layers which are separated by a thi n dielectric layer. The upper metallic layer is responsible for

PAGE 83

83 radiation while the lower one works as the ground plane blocking EM waves propagating to the ground plane direction. Since the patch antenna shows good directivity in the broad side of the upp er patch, electronic circuitry placed behind the ground plane would have little EMI from the patch antenna radiation. Usually an omnidirectional pattern is obtained by using an omnidirectional radiator such as a monopole or dipole antenna but those antenna types are huge in footprint and will generate EM interference to the surrounding circuits. Another way is to manipulate the antenna shape to reconfigure the directional pattern into an omnidirectional pattern. In this work, we simply wrap a planar patch a ntenna into a three dimensional (3D) rectangular waveguide shape to achieve all the requirements listed above including an omnidirectional pattern. The antenna design begins with a conventional patch antenna as shown in Figure 4 3 A where the substrate has a dielectric constant of 2.33, a thickness of 31 mil s (0.8 mm) An approximate patch length L can be calcul ated by the following equation: ( 4 2 ) and is the relative permittivity of the substrate. The conventional antenna length (z direction) is determined to be 40 mm for a resonance radiation frequency of 2.4 GHz, and the antenna width (x direction) is chosen to be 70 mm. A the feeding line for impedance matching. Then, this conventional patch is folded into a rectangular waveguide shape as shown in Figure 4 3 B This antenna has t he same antenna length of 40 mm (z direction) and the resultant waveguide structure has a dimension of 20 mm x 20 mm in its cross section, much reducing its footprint. The folded patch antenna does not have both edges meet each other but has a gap between both

PAGE 84

84 edge s This gap (or slot) can be further cut away, to satisfy the impedance matching, not necessitating the This folded patch antenna has been simulated using HFSS. Figure 4 4 A shows that the maximum return losses of the conventional patch and the folded patch are 25.4 dB at 2.41 GHz and 35.4 dB at 2.38 GHz, respectively. The folded antenna is fabricated using a Rogers RT5870 substrate whose permittivity and thickness are 2.33 and 30 mil s respectively, as shown in Figure 4 5. Measurements are performed using an HP 8510C vector network analyzer after a standard 1 port calibration between 1 GHz and 4 GHz. Figure 4 4 B shows that the measured return losses of the conventional and folded patch antennas are 32.2 dB at 2.4GHz and 53 dB at 2.41 GHz, respectively, showing a very good impedance matching for the folded patch antenna. In order to test an EMI effect, we fill th e internal space of the folded antenna with pieces of printing circuit boards (PCBs) and measure the return loss. The resonance frequency remains the same at 2.41GHz while the maximum return loss changes to 27.2 dB. The o mni directional pattern is obtained for both cases. 4.3 Compact Wrapped Patch Antennas It can be seen that the breakthrough of the 3D folded patch antenna is the achievement of the omnidirectional pattern with EMI shielding and via free easy integration. However, it is obviously too huge for mo st of the medical applications. In this section, we will investigate several smaller versions of the 3D folded patch antennas. 4.3.1 By Loading S hor ting V ias As shown in Figure 4 7 A the E field distribution on a patch antenna has a null in the patch center wher e a virtual ground is expected. So the most straight forward way to reduce the antenna size to half is to load shoring vias in the center of the patch.

PAGE 85

85 After careful tuning, the vertical length of the patch is determined to be 19 mm, the cross section of the antenna has an area of 20 mm 20 mm, and electric vias are put by the edge Also, inductive dents are designed and optimized which are put on the back side of the antenna to compensate the inductive input impedance due to size reduction. Comparing it with the previous version in the previous section, the proposed antenna has achieved a size reduction up to 52.5%. The p roposed antenna is fabricated based on the simulation results. We are using Rogers RT/Duroid 5870 as the substrate for fabrication which has a relative permittivity of 2.33, a thickness of 31 mil s The fabricated antenna is shown in Figure 4 7. The a ntenna is measured on an HP 8719D network analyzer after a standard 1 port calibration. The measured result is also included in Figure 4 8 wi th simulation data for compar ison The measured resonance frequency is at 2.38 GHz, with a return loss of 29.3dB. 4.3.2 By M etamaterial Loading on The Back G round CSRR resonance frequency is given by ( 4 3 ) D ue to duality, and can be found from its corresponding SRR by [39]: (4 4 ) (4 5 ) where as shown in Figure 4 10 is the average radius of the CSRR rings; is the per unit length capacitance between the rings Ls can be approximated by that of a single ring with averaged radius and width [39]. And Based on E quations ( 4 3 ) through (4 5 ), on a substrate with a relative permittivity of 2.33, a thickness of 31 mil s to have a CSRR resonance at 2.4GHz, a radius of the outer ring rout = 7.5 mm, the distance between traces d = 1 mm, and the trace width t = 1 mm.

PAGE 86

86 In this work, the proposed CSRR will be placed by the edge of the ground plane of the folded patch antenna. The CSRR structure effectively reconfigures the electromagnetic field of the antenna, and the vertical length is reduced to 50% of its original size that is 20 mm on the same substrate. Meanwhile, the cross section area has been also reduced to 15 mm 15 mm. The patch does not cover all around the surface of this structure but there are three slots on the patch where the one on the back is 5 mm 1 5 mm in size while the ones on the front are 2 mm 11 mm. All those three slots are important for input impedance tuning of the antenna. The CSRR outer ring has both the trace width and the distance between traces of 1 mm. The CSRR edge is aligned with th e upper edge of the patch. In simulation, the antenna radiates at 2.4 GHz with a maximu m return loss (S11) of 10.27 dB This antenna offers a nearly omni directional radiation pattern while the maximum directivity is as low as 0.2 dBi. The r adiation effici ency is 60%. It has been shown that with careful tuning, the antenna radiates at the same center frequency as one from the previous work while the volume is greatly reduced by 72 %. It is also shown that this newly proposed antenna has a quasi omni directi onal radiation pattern which is desirable in many applications such as biomedical systems and compact sensors while it provides an EMI shielded cavity useful for other electronics installation. This kind of radiation pattern is very difficult to obtain fro m the traditional patch antenna s The p roposed antenna is fabricated on Rogers RT/Duroid 5870 (relative permittivity = 2.33, thickness = 31 mil) using a milling machine. Since the patch and the CSRR are on the different sides of the PCB, a double layer fab rication process and careful alignment are required. The fabricated device is shown in Figure 4 12 It is further demonstrated that in Figure 4 12 C that this CSRR is located inside the cavity of the folded patch antenna and in a non planar shape.

PAGE 87

87 Measureme nts are performed on an HP 8719D network analyzer after an S11 1 port calibration. The measured results are compared with simulation ones in Figure 4 11. The measurement data show the radiation frequency at 2.36 GHz with a return loss of 18.4 dB. There is a frequency shift which is attributed to the mis alignment of the CSRR. 4.3.3 By Metamaterial L oading on The Patch S urface F rom the equivalent circuit point of view, the additiona l CSRR is to add another resonator onto the original patch antenna for size reduction. The advantage of using the CSRR lies in that the resonance frequency of CSRR does not rely on its physical size but its equivalent circuit model. In this section, the antenna is fabricated on a flexible and thin substrate, a CSRR particle is loaded on the surface of the patch by which the antenna size is greatly reduced. The p atch length is reduced to 0.11 e first 3D folded patch antenna. Further, this antenna is fabricated on a flexible substrate Now the antenna is in a real cylindrical capsule shape which is more appropriate to be used in biomedical applications such as in a wireless capsule endoscope. M eanwhile, the CSRR is etched on the surface of the patch which is easier to im plement However, since the antenna size is too small and the CSRR shows high Q the antenna demonstrated here has a low antenna gain and a low radiation efficiency. Different f rom earlier work, here a folded patch antenna with CSRR loading is designed on a flexible substrate, RT/Duroid 5880 (Rogers Inc.) with a dielectric thickness of 10 mil s (0.254mm), a clad copper thickness of 17 m, a dielectric constant of 2.2, and a loss tangent of 0.0009. By accommodating the flexible substrate, the folded shape can be much diversified e.g. a cylindrical shape, which is more ideal for capsule applications, is feasible. Also, the CSRR

PAGE 88

88 structure i s patterned in the patch to reduce the size of the antenna. Because of the folded architecture, the resultant CSRR is in a non planar shape. Its design and analysis is performed using circuit modeling and HFSS simulation A patch antenna folded in a 3 D cy lindrical shape ( Figure 4 13 A ) is modeled as a RLC resonator tank as shown in Figure 4 13 B The dimensions and the extracted lumped element values are given in the caption of Figure 4 13. L f stands for the inductance value of the feeding line and the RLC r esonance tank represents the patch. It should be noticed that after folding, there is a gap (or slot) between the two edges of the patch on the opposite side of the feeding line, whose dimension is denoted as g in Figure 4 13 A This slot can be used for im pedance matching and therefore no external matching circuit is necessary. HFSS simulation for the structure of Figure 4 13 A shows its resonance frequency of 9.16 GHz and the equivalent circuit of Figure 4 13 B presents a resonance frequency of 9.155 GHz. Th e patch antenna shown in Figure 4 14 A is the combination of a CSRR and the patch antenna of Figure 4 13 A with the same dimensional parameters except two additional matching slots at the base of the patch which have a size of 2.99 mm by 4 mm. The CSRR is lo aded on the top portion of the patch antenna. The overall system is represented with two resonator tanks (one from the plane patch and the other from CSRR) connected by LC couplings as shown in Figure 4 14 B and the whole system radiates at a balanced freq uency of 2.4GHz. The antenna length f is 10.5 mm (0.11 compared with its traditional patch counterpart without CSRR loading. The simulated antenna has a front to back ratio as l ow as 0.5 dB which indicates a quasi omnidirectional pattern. Since

PAGE 89

89 the antenna is electrically small, it shows a relatively low gain of 5.2 dB. The antenna radiation efficiency is 21 % The antenna serves as an electromagnetic radiator and possibly a packaging layer The antenna has no electromagnetic leakage toward the enclosed area where EMI shielding is guaranteed by the ground plane of the patch antenna. High speed digital circuits and EMI susceptible analog circuits can be loaded inside the antenna without interference. A ntenna optimization with a more realistic human body model would be p erformed for which the patch dimensions and matching slots need to be modified. 4.4 Multifunctional Wrapped Helix Antenna Conventional wired endoscopes, due to a limited reachable span and possible trauma associated with the attached wire, are being replaced by capsule endoscopes. Since the capsule relies on wireless link for data communication, an efficient antenna is a key component. Also, important is its power source to operate a light, a camera, various sensors, and electronics. With a space constraint, h owever, a few button cells are used and they won't last long for exhaustive endoscope and communication activities through the lossy body media. Meantime, inductively coupled RF power can be considered as an alternative power source. So an advanced capsule endoscope requires efficient antennas for wireless data communication and power transmission. In this work, a dual function helix shape structure is presented, where it functions as an RF antenna for wireless data communication at 400 MHz and an inductor for near field wireless power transmission to charge the capsule at 150 kHz. The helix is formed by rolling up a flexible substrate with a narrow copper trace into a cylindrical shape by which an omnidirectional radiation pattern, desirable to reduce commu nication blind spots, is obtained. Good radiation efficiency prolongs battery life. The rolled up antenna layer serves as a protection/packaging

PAGE 90

90 layer for the capsule. The helix also functions as a charging coil for near field power transmission. The desig n, fabrication, characterization, and dual functionalities of the helix are demonstrated. The p rinciple of operation is that the high frequency (400 MHz) signal goes to the impedance matched path for wireless communication and the low frequency ( 150 kHz ) g oes to the power management circuit for wireless charging. 400 MHz does not go to the charging path because of impedance mismatch while 150 kHz does not go to the communication path because of the band pass filter. The antenna is designed on a Rogers RO300 3 substrate with a dielectric constant of 3.0 and a thickness of 10 mils. As shown in Figure 4 17 the antenna is composed of a piece of microstrip feeding line with a width of 0.65 mm and a titled trace with a length of 220 mm and a width of 0.3 mm. A tilting angle of 1.9 is used for the desired helix with a pitch of 0.63 mm. Due to the flexibility of the substrate, the antenna can be wrapped into a cylinder with a desired diameter of 6 mm. The antenna is fabricated using a milling machine S100 from LP KF Inc. The fabricated device is shown in Figure 4 17. Far field measurement is performed using an Agilent E5071C vector network analyzer. As shown in Figure 4 18, a good radiation pattern is demonstrated at 400 MHz with a bandwidth of 21 MHz, encompassing the MICS band (402~405 MHz). An omnidirectional radiation pattern is obtained with an antenna gain of 10 dB and a radiation efficiency of 89.4 %. At 150 kHz, the inductance of the helix is 548 nH and the stray A power management IC fo r near field wireless charging demonstration is designed using the 0.5 m CMOS technology. Figure 4 19 shows DC outputs for different levels of AC input s showing a stable DC voltage output. The proposed antenna is designed and demonstrated

PAGE 91

91 for data communi cation and wireless charging for a wireless endoscope system. However, the concept can be applied to oth er autonomous wireless systems. 4.5 Super Compact Wrapped Patch Antenna with Inductive Slot Loading In this work, a compact patch antenna with multiple indu ctive notches loading is designed in a human body model environment, fabricated on a flexible substrate, and wrapped into a cylindrical shape for wireless capsule endoscope applications. The antenna is designed to have a radiation frequency of 433 MHz and an omnidirectional radiation pattern is obtained after wrapping it into the cylindrical shape. Human phantom solution is made according to FCC specifications and the antenna is characterized in the solution. Since the wireless endoscope antenna operates in a lossy and dispersive human body environment, it is desirable to have human body effects taken into consideration in antenna design. A real tissue environment could be utilized for antenna design and characterization. However, it is not an economic and p ractical approach. Another way is to approximate the human body as a homogeneous and linear medium whose constitutive parameters are close to those of the human body. Although this method offers a simple solution, it does not consider the frequency depende nt nature of the human body tissues and the accuracy is subject to doubt especially for precision medical applications. In this work, a numerical human body model of the high frequency structure simulator (HFSS, Ansys Inc.) is used to assist antenna design and analysis. The human body model offers frequency dependent microwave constitutive parameters including electrical permittivity and conductivity for most of human organs and tissues at different accuracy levels in male and female versions. The average m icrowave properties of such a model can be extracted as shown in Figure 4 20.

PAGE 92

92 Figure 4 20 shows the frequency dependency of the human body constitutive parameters. Antenna performance including the radiation pattern, radiation frequency and antenna gain is all characterized based on this frequency dependent human body model for better accuracy and design efficiency. Slits or notches loaded on a microstrip component can be analyzed by the method of perturbation. As shown in Figure 4 21 the notch leads to a localized concentration of magnetic field, and thus it can be treated as an inductive loading. Since the microstrip width changes at the notch, characteristic impedance variation takes place from Z 0 to 0 and the equivalent inductance for each notch can be calculated from Eq uation ( 4 6 ): ( 4 6 ) where is the magnetic permeability of free space, h is the thickness of the substrate [ 40 ]. It should be noticed that the equivalent inductance introduced by the notch is independent of the substrate permittivity [ 40 ]. A patch antenna fabricated on a flexible substrate can be wrapped into a cylindrical capsule shape. In this work, the antenna is designed on a liquid crystalline polymer (LCP) flexible substrate with a thickness of 4 mil s (0.1 mm) and a dielectric constant of 2.9. The antenna is designed using the HFSS ( Ansys Inc.), where the human body model is used as a surrounding environment. Two radiation slots are modeled by its radi ation admittance Y = G + j C, where G is the shunt conductance and C is the shunt capacitance. Since those slots have the same physical size and if we ignore the fringing effects of the feeding line, we have Y 1 =Y 2 The inductance component L comes from th e feeding line as well as the patch. A patch antenna with a center radiation frequency of 433 MHz is targeted.

PAGE 93

93 A patch that has no notch loading is shown in Figure 4 22 The cylinder dimensions are designed to be comparable with those of the commercially available wireless endoscope capsules. It can be seen in Figure 4 22 A that a Polydimethylsiloxane (PDMS) layer with a thickness of 0.5 mm is coated on the surface of the antenna. PDMS is optically transparent in the visible light and bio compatible. The transparent PDMS allows an integrated camera to record the inside images. With a patch width of 18.5 mm and a capsule diameter of 1 0 mm, the resonant radiation frequency can be determined by: (4 7 ) where C reflects all the equivalent capacitances associated with C 1 and C 2 as shown in Figure 4 22 B HFSS and equivalent circuit model simulation shows that t his patch antenna radiates in resonance at 1.38 GHz, which is higher than the targeted frequency of 433MHz. From Eq uation ( 4 7 ), we can see t o lower the resonance frequency with proper inductive or capacitive loading S ince capacitive loading usually requires additional components or an extra area, in this work, multiple notches with a dimension of each notch of 9.15 mm by 0. 25 mm are inserted for inductive loading. The notches are implemented by subtracting some areas from the patch and the overall size is not increased. The overall patch capacitance may be reduced but if we increase the inductance value significantly more, the resultant resonant radiation frequency can be reduced down to the desired frequency. According to Eq uation ( 4 6 ), the inductance of each pair of notches is about 0.1 nH. To bring down the resonance frequency of the antenna to 433 MHz, 39 pairs of notches are loaded on the patch. The geometry of the inductively loaded patch is shown in Figure 4 23

PAGE 94

94 It should be noted that those notches are also helpful to make it easy to fold the antenna into a cylindrical capsule shape. The antenna shown in Figure 4 23 B can be modeled as a n L C tank with a series of inductors loaded. The equivalent circuit model is shown in Figure 4 23 C Circuit model simulation shows the same radiation frequency as the HFSS electromagnetic simulation. The antenna is fabricated on the flexible substrate using a milling machine (S100, LPKF Inc.). The fabricated antenna is shown in Figure 4 24 As shown in Figure 4 24 B the feeding line is wrapped inside the cylinder cavity, and the SMA connector is used for testing. The fabricated antenna is tested in a human phantom whose composition is shown in Table 4 1. Measurement is performed using an Agilent E5071C vector network analyzer after one port calibration. A bare antenna, the PDMS coated antenna in air, and the PDMS coated antenna in the phantom solution have been tested. This phantom solution has a dielectric constant of 58.00 and a conductivity of 0.83 S/m at 433 MHz which is very close to the human body model we have used in simulation with a dielectr ic constant of 58.11 and a conductivity of 0.83 S/m in the same band (refer to Figure 4 20 ). The measurement setup and the S parameters are shown in Figure 4 25 The bare antenna radiates at 540 MHz while the PDMS coated one radiates at 436 MHz. In the hum an phantom solution, the antenna shows radiation at 430 MHz with a peak return loss of 18 dB. The frequency shift from the designed 433 MHz to 430 MHz is attributed to the fabrication tolerance. The antenna is grabbed by a human hand, which shows nearly th e same result as the one in the human phantom solution in terms of the radiation frequency. Note that no external matching circuit is used for this antenna. With inductive loading, the patch length is

PAGE 95

95 only approximately 0.0 7 g, where g is the guided wave length on the substrate. It should be noticed that the human body effect is suppressed when PDMS coating is placed on the antenna. Due to the electrically small size, t he antenna gain is relatively low, and is 9.6 dBi. However, the transceiver (CC1110, Texas Instruments) we use for the system has a sensitivity of 110 dBm, and so this antenna gain would be acceptable. Antenna radiation efficiency is 49% in free space but drops to 1.9% in the human body phantom, which is attributed to the absorp tion nature of the electromagnetic waves in the human body. The antenna shows an omnidirectional radiation pattern in the free space environment as shown in Figure 4 26 A In the human body model environment, the omnidirectional radiation pattern is degrade d as shown in Figure 4 26 B Constitutive parameters vary from organ to organ, and thus the degradation of omnidirectionality is due to the asymmetric dielectric and conductivity distribution in the human body. Human movement also has an effect on the radia tion pattern. In Figure 4 27 A the human body has both arms placed by the body sides. W hen the human body raises both arms by 40 degree as shown in Figure 4 27 B the radiation pattern is further changed as shown in Figure 4 28 With both arms rising up the radiation pattern on the x y plane (azimuth plane) spreads out toward the direction of the raised arms by about 2 dB on both sides. Radiation frequency also shifts to a lower band by about 1.75 MHz when arms are raised. Since the front and back side of the body has stronger radiation signals, the outside receiver can be placed on the front or back rather than the sides in order to have better signal reception. A n inductive ly loaded folded patch antenna has been demonstrated for a wireless endoscope application. The antenna is designed on a flexible LCP based substrate and wrapped

PAGE 96

96 into the desired capsule shape T his antenna is compact and shows an omnidirectional radiation pattern. A ntenna characterization is performed in the human body phantom solution and the HFSS human body model. Human movement effects on the antenna performance are studied The wrapped antenna offers a novel circuit assembly approach. A ll necessary sensors and transceiver electronics can be designed on the sam e plane with the antenna on a flexible substrate. T he demonstrated rolling up approach exercises easy assembly and would reduce the design and implement ation cost. T h e antenna offers mechanical protection packaging, and EMI shielding to the high frequency electronics in the capsule.

PAGE 97

97 Table 4 1 Composition of a h uman body phantom solution Item Percentage by weight Deionized water 51.16 Sodium Chloride 1.49 Sugar 46.78 Bactericide 0.05 Hydroxyethyl 0.52 Figure 4 1 Microstrip patch antenna schematic A

PAGE 98

98 B C Figure 4 2 Field Distribution on a patch antenna A) E field. B) Surface current density. C ) H Field

PAGE 99

99 A B Figure 4 3 Patch antennas. A ) Conventional planar patch antenna B ) Proposed folded patch antenna i n a rectangular waveguide shape A B Figure 4 4 Conventional and folded patch antenna performance A ) Simulation B ) Measurement A B Figure 4 5 Folded patch antenna in a rectangular waveguide shape A ) Feeding line side B ) Back side with a gap (or slot) -40 -30 -20 -10 0 1 1.5 2 2.5 3 3.5 4 Conventional Folded Frequency/GHz S11/dB -60 -50 -40 -30 -20 -10 0 1 1.5 2 2.5 3 3.5 4 Conventional Folded Frequency/GHz S11/dB z x y Fold ed 40mm 40mm Feeding Line Tuning Gap (Slot)

PAGE 100

100 Figure 4 6 Performance of a folded patch antenna with PCB filling A B C D Figure 4 7 Compact omnidirectional folded patch antenna A ) Schematic, front B ) Schematic, back C ) Fabricated device, front D ) Fabricated device, side view Figure 4 8 Measured and simulated return loss -60 -50 -40 -30 -20 -10 0 1 1.5 2 2.5 3 3.5 4 No Filling Filled with PCB Frequency/GHz S11/dB -35 -30 -25 -20 -15 -10 -5 0 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 Measured Simulated Frequency /GHz Reflection Coefficient (S11) /dB

PAGE 101

101 A B Figure 4 9 Radiation pattern A ) H plane B ) E plan e A B Figure 4 10 Folded patch antenna with CSRR loaded on the ground plane A ) Before folding B) Folded Figure 4 11 Return loss -20 -16 -12 -8 -4 0 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 Measurement Simulation

PAGE 102

102 A B C Figure 4 12 Fabricated antenna A ) Front B ) Side C ) CSRR on the ground A B Figure 4 13 Patch antenna folded in a cylindrical shape A ) Schematic B ) Equivalent circuit. Dimensions: f = 10.50mm; g = 2.88mm; h = 10.00 mm. Extracted lumped elements: A B Figure 4 14 CSRR loaded patch antenna A ) Schematic B ) Comprehensive equivale nt circuit model. Extracted lumped elements: Lc1 = 4.57nH; Cc1 = 1.35pF; Lc2 = 1.08nH; Cc2

PAGE 103

103 Figure 4 15 Fabricated antenna and measured results A B C Figure 4 16 Schematic of the proposed helix antenna A ) Metal trace on a flat substrate B ) Wrapping of the metal trace C ) Fully wrapped helix antenna . A B C Figure 4 17 Fabricated Antenna A ) Front side B ) Back side C ) System schematic -16 -14 -12 -10 -8 -6 -4 -2 0 1.5 1.7 1.9 2.1 2.3 2.5 2.7 2.9 HFSS Simulation Circuit Simulation Measured, no sponge Frequency / GHz Reflection Coefficients ( |S11|)/ dB

PAGE 104

104 A B Fig ure 4 18 Antenna far field performance A ) Reflection coefficients B ) Omnidirectional pattern A B Figure 4 19 Near field charging performance A ) Measured DC output at different AC input levels B ) Optical photo of the power management IC Figure 4 20 Frequency dependent human body parameters. Inset: Ansys HFSS male human body exterior -40 -35 -30 -25 -20 -15 -10 -5 0 100 200 300 400 500 600 700 Frequency [MHz] LogMag (S11) [dB] 0 0.5 1 1.5 2 2.5 3 4 5 6 7 8 9 Regulated DC Output Voltage [V] Input AC Peak Voltage [V] 0 1 2 3 4 5 6 0 20 40 60 80 100 120 140 160 180 200 100 200 300 400 500 600 700 Relative Permittivity Conductivity[S/m] Frequency [MHz] Conductivity [S/m] Relative Permittiivty

PAGE 105

105 A B Figure 4 21 Inductive notch loading on a microstrip component A ) Top view of the notch B ) Equivalent circuit A B Figure 4 22 Wrapped patch antenna without notch loading A ) Schematic B ) Equivalent circuit model: A B C Figure 4 23 Wrappable patch antenna with inductive loading A ) Stretched flat patch B ) Wrapped patch C ) Equivalent circuit of the wrapped patch with inductive loading: = 0.1 nH x y z

PAGE 106

106 A B Figure 4 24 Fabricated patch antenna with inductive loading on a flexible substrate A ) Before PDMS coating B ) After PDMS coating Figure 4 25 Return loss of the fabricated antenna under different conditions Inset: measurement setup (the antenna is inserted in the human phantom solution.) Coated in air means the antenna with PDMS coating and measurement i s done in free space; PDMS coated in hand means the coated antenna is measured with human hand surrounded; PDMS coated in human phantom means the coated antenna measured in the human phantom solution A B -30 -25 -20 -15 -10 -5 0 100 200 300 400 500 600 700 Bare Antenna Coated in Air PDMS Coated in Human Hand PDMS Coated in Human Phantom Frequency [MHz] LogMag (S11) [dB] -30 -25 -20 -15 -10 -43 -33 -23

PAGE 107

107 Figure 4 26 Antenna radiation pattern A ) In free space B ) In human body model. Solid line: x z plane; dashed line: x y plane A B Figure 4 27 3D radiation pattern with human body model A ) Both arms placed by the body sides B ) Both arms raised up by 40 degree A B Figure 4 28 Radiation pattern change according to arm positions A ) z x plane B ) x y plane. Dashed line shows the case with arms raised and solid line shows the case with arms by the body sides -43 -38 -33 -28 -23 -43 -33 -23 x y z

PAGE 108

108 CHAPTER 5 WIRELESS CAPSULE ENDOSCOPE 5.1 System Scopes and Radio Specifications Endoscopic diagnosis is of great importance in gastrointestinal (GI) tract disease diagnosis, as well as early tumor treatment [4 1 ]. However, their depth of insertion limits performance and diagnosis capability of the traditional wired endoscopes Meanwhile, bleeding lesion associated with endoscopic diagnosis has been reported in [4 1 4 2 ]. Wireless endoscopy is now a popular research area where the communication and power cord are replaced by wireless channel. First wireless endoscopic diagnosis wa s demonstrated as early as 2000 [4 3 ] Wireless capsule endoscopy avoids the risk of sedation and intubation associated with conventional endoscopic methods A lternatively, people use radiological investigation for small intestine imaging, in which case patients need to be exposed to radiation. The d esign challeng e of a wireles s capsule endoscope mainly lies in the tradeoff between the system performance requirements and the limited resource available. Design constraints include: 1) Mechanical: The most significant problem for endoscopic diagnosis lies in the po ssible retention of the capsule [4 5 ] The physical size of the capsule should be reasonable. Commercially available capsule endoscopes usually have a dimension of 10 mm by 20 mm (diameter by height) which is a good reference size. Meanwhile the system should have sufficient mechanical supports as well as robust packaging for the system. 2) Radio performance: As we have learnt from Chapter 1 400 MHz, 900 MHz and 2.4 GHz are three of the most popular license free band s for medical application s. Among them, 400 MHz have the minimum path loss and thus an optimized frequency for best system performance

PAGE 109

109 (lower BER, larger detection range) while 2.4 GHz has the minimum wavelength which assists device miniaturization. In this work, we begin with a 2.4 GHz system which has a smaller size by nature, then we use certain advanced technologies to implement a system with a similar size but operating at the 400 MHz band. The system should have enough bandwidth for smooth information delivery, which require s the RF transceiver, signal conditioning circuits and the antenna with a certain bandwidth. A wireless s ystem should choose a proper modulation scheme so that the desired signal to noise ratio ( SNR ) require ment is met while keeping low energy consumption from the power source. Frequency modulation such as Frequency Shift Keying ( FSK ) eases the microprocessor load A s we have discussed before, since it has a constant signal envelope, it allows the amplifier operate near the 1 dB compression point where maxi mum power efficiency is obtained, so that FSK would be the most power efficient modulation scheme. However, since FSK is a frequency multiplying method, the carrier frequency occupies a relatively wide frequency range, and the FSK requires wider bandwidth comparing to other modulation methods. Instead we may use Phase Shift Keying ( PSK ) which requires less bandwidth but due to its modulation complexity, the PSK system requires more processor efforts M eanwhile, PSK systems do re quire amplifier work in the linear region, and the power efficiency is not maximized for this kind of modulation. Other modulations are usually too complicated to be implemented in such a budget tight system. 3) System cost: Overall system cost of a capsu le endoscope should be kept within a reasonable range. In traditional wireless endoscope systems, great design efforts are needed for mechanical supporting between layers and electrical interconnection [4 6 ] In this work, a novel design would be proposed t o simplify those structures while maintaining good performance.

PAGE 110

110 4)Performance: I n traditional endoscope systems, interconnections between subsystems are usually difficult to design E specially for RF transceiver subsystem s the interlayer connection vias a re lossy and a significant insertion loss is introduced by such lossy vias. 5) Power consumption: P ower supply has been a major concern for a wireless endoscope. The most power hungry part on a wireless endoscope is the RF subsystem. As f or an antenna, to minimize the power waste, we need to make sure the antenna has proper impedance matching. For an electrically small antenna it is usually very difficult to match with the desired characteristic impedance T hose small antennas are usually small in input i mpedance but huge in input reactance, and a proper matching design is usually a must. 5.2 Planar Patch Antenna with Inductive Loading Electrically small antenna is preferred in modern wireless medical applications such as in a wireless endoscope. From the discussion in Chapter 4, when a microstrip patch antenna is forced to work at a relatively low frequency (i.e. the antenna size is mad e much smaller comparing to the wavelength), there are several concerns to be addressed. First, antenna impedance matching would be very difficult. For electrically small antenna s the real part of its input impedance approaches zero and the imaginary part becomes negative which means capacitive input impedance. Second, antenna efficiency is reduced. And according to Eq uation ( 1 2) in Chapter 1, with smaller aperture, the antenna gain degrades. Impedance matching is the most significant problem needs to be addressed. Real impedance mismatching can be alleviated by tuning the patch and length ratio or using a quarter wavelength transformers H owever, huge negative imaginary input impedance is difficult to tune. In this work, a similar method as shown in Secti on4.5 is used. Inductive microstrip components are loaded to compensate the capacitive impedance

PAGE 111

111 5.3 Capsule Circuitry Design and Antenna Integration Due to the flexibility of the antenna as we have studied in the previous section, a capsule can be built by integrating all necessary electronics as well as the antenna onto a single piece of flexible substrate, and then the capsule can be obtained by wrapping the substrate into the desired cylindrical shape. For demonstration purposes a wrapped capsule with a temperature sensor working at 2.4 GHz is demonstrated. It should be noticed that using the similar technology we have discussed in Section 5.2, a patch antenna with inductive loadings is used for this system. The antenna dimension is shown in Figure 5 1 The antenna is designed on a n LCP based substrate ( Rogers Ultralam 3850 ) which we have studied in Chapter 2, with a thickness of 4 mil s a dielectric constant of 2.9 and a loss tangent of 0.0012. The antenna is fabricated using a foundry service. The antenna is characterized using an Agilent E5071C network analyzer after calibration. The return loss results are shown in Figure 5 2 It can be seen that no matter the patch is flattened or rolled up, the measured radiation frequency r emains as designed at 2.41 GHz. The simulated antenna radiation efficiency is 60.2% and the antenna gain is 0.02 dB. The a ntenna has a bandwidth of 100 M Hz The antenna shows a reasonable omnidirectional pattern as shown in Figure 5 3 It should be noticed that after wrapping, the cavity enclosed by the patch antenna is shielded by the ground plane of the patch, and thus an EMI free cavity zone is generated. The antenna does not require any additional matching circuits to interface with a 50 o hm system, since the slots and the gap between two ends can be used for impedance matching, making the antenna compact. The g ood impedance matching and the relatively high radiation efficiency contribute to the reduction of power consumption of the w hole system.

PAGE 112

112 The implemented wireless endoscope platform is composed of a capsule shape endoscope, a USB type receiver, and a computer program for data acquisition as well as system control. The communication protocol used between transceivers is SimpliciT I by Texas Instruments (TI) Inc. The wireless node is based on CC2510 from TI which is a low power transceiver integrated with a 8051 core microprocessor. The transceiver and other sensors are implemented on the same flexible substrate with the wrappable patch antenna. The system is shown in Figure 5 4 The implemented system does not require complicated and delicate multiple layer stacking up design and complex signal integrity analysis. The antenna provides omnidirectionality as well as EMI protection to the system. The system is powered up by two coin cells (392 384TZ, Energizer Inc.). 5.4 System Controller and Communication Protocol Design The collected data are then received by a USB type receiver and displayed on a PC. The receiver is shown in Figure 5 5 B this dongle is based on TI eZ430 CC2510 platform, the developed graphic user interface is shown in Figure 5 5A where both are based on TI MSP 430. The communication protocol is based on SimpliciTI from TI which is a low power protocol for small scale low data rate networks. We used 2 FSK as the modulation scheme for best power efficiency and the required system bandwidth is 103 kHz. However, since ou r antenna bandwidth is about 100 M Hz, we may los e part of the data packages which may lead to high BER. The carrier center frequency is 2410 MHz, and all those settings result in a data rate of 2.4 kBaud. 5.5 Hybrid Circuitry Design Flexible circuit has its un ique advantages H owever, the cost to assembl e a piece of flexible system is high. Special fixture is a must to mount parts precisely on a flexible substrate.

PAGE 113

113 Another drawback of a purely flexible system is the very weak mechanical robustness. The soldered parts are very easy to peel off from the hosting substrate. We may smartly combine the advantage of flexible circuits and the traditional rigid ci rcuits. As shown in Figure 5 6 instead of using whole flexible circuits, we move the sensor and camera circ uits to a piece of rigid circuits while keep ing the antenna on a flexible substrate. The flexible antenna used in this hybrid version of wireless endoscope is introduced in Section 4.5, e.g. a super compact wrapped path antenna with inductive loading. As shown in Figure 5 6 C, the wrapped antenna still works as the outmost layer of the capsule, the rigid controlling circuits are enclosed in the flexible antenna. All the electronic components are assembled on the rigid section, thus there is no necessity to do components mounting on the flexible circuit, and the assembly cost is reduced It should be noticed that although we use rigid circuit components and boards we do not have an antenna on the rigid board, so that the circuit design is still simple. In this system, a normal double layer board is sufficient. However, the antenna is still designed and fabricated on the flexible substrate as described in Section 4.5, the antenna is still wrappa ble and the omnidirectionality is reserved. Since we still have the ground plane surrounding the rigid circuits, EMI protection is still preserved for the antenna structure. The hybrid circuit implementation combines the low cost of rigid circuit while re serving the EMI and omnidirectionality of a wrappable antenna circuit on a flexible substrate Performance and cost is well balanced from this novel configuration. 5.6 Sensor and Camera Implementation We make use of a temperature sensor integrated on the ADC t o monitor temperature. Real time temperature with a resolution of 0.1 0.5 wireless sensor measures temperature of 180 mL water. The water is contained in a beaker which

PAGE 114

114 is heated up by a hot plate with a temperat ure ramping rate of 400 Scientific Inc.) is used as a reference. The wireless sensor system is protected with water proof coating. The measured real time temperature can be read as shown in Figure 5 6A and the measured results are shown in Figure 5 7A The discrepancy between the two measurement results is mainly due to the coating of the wireless sensor which has lower heat conductivity. As shown in Figure 5 7B, the core section of this system is CC1110 from TI. This is a low power SoC with a sub 1GHz wireless radio transceiver and an enhanced 8051 compatible processor core. This SoC needs very intensive decoupling design for both signal integrity and power integrity purposes. There are two sensors implemented in the system. T he integrated power sensor in CC1110 measures the temperature. The pressure is measured by MPL115A2 from Freescale Inc. This IC includes a MEMS pressure sensor and related signal conditioning circuitry. I 2 C bus is used to communicate between this sensor and the SoC. Our system offers pressure measurement capability of 7kPa. As shown in Figure 5 7C, since the I 2 C bus has an open drain design, pull up resistors are needed. To extend battery life, we implemented a power conditioning circuit. A boost DC/DC c onverter TPS61221 by Texas Instrument Inc. is used. After running for a while, battery voltage drops below certain value, after this threshold, the system powers off due to insufficient voltage value. Our power module is carefully designed so that the syste m is still running even the battery voltage has dropped below 0.8V. However, since this power module is a high frequency and high power module, it should be placed further away and well isolated from the radio frequency circuit and antenna. In addition to multiple sensors, a compact camera OVM7690 from OmniVision is used for image acquisition as shown in Figure 5 7E Comparing to other cameras, OVM7690 allows a relatively wide input voltage range (2.6 3.3V) I t draws a peak current of 40 mA, and a SCCB

PAGE 115

115 b us which is similar to that of I 2 C bus which makes it easy for the microprocessor to commu nicate with the camera module [ 48 ] T he capsule endoscope block diagram is shown in Figure 5 8. Since many parts of human gastrointestinal tract is very n arrow for wireless endoscope, it is critical to make sure that the camera can focus when the sho o ting distance is very short. As shown in Figure 5 9 A to C we use a 100RMB Chinese currency note as an object W hen the distance is above 2.45 cm, the image quality is acceptable H owever, if the object distance is euqal or smaller than 1cm as shown in Figure 5 9 C, the image is blurred. This problem is alleviated by borrowing a new piece of sample code from the camera vendor. Although the vendor does not reveal the detail principle the near distance focus capability is greatly improved. Even with a 1cm distance we can have a clear image as shown in Figure 5 9 D. Four light emitting diodes ( LEDs ) are also integrated to assist image acquisition. LNJ047X8ARA from Pana sonic electronics is used. Those LEDs have a dimension of 1 mm x 0.5 mm x 0.2 mm with a forward current of 5mA and a luminous intensity of 50 mcd In software we set the camera to turn on and take a snapshot every 5 seconds. Accordingly, LEDs flash ever y 5 seconds. In between snapshots, the whole system is put in a sleep mode, for which much less than 1mA of current is consumed Different from the previous version, the hybrid wireless endoscope uses a wireless receiver to collect the data including images. The idea is to let the patient carries the data collector after swallowing the endoscope. The data collector is shown in Figur e 5 10.

PAGE 116

116 The receiver is composed of a mother board, which controls the digital IO and offers some peripheral functionalities such as display and voice playback. We have a CC1110 SoC installed on the board. A n SD card module has two operation mode: Serial P ort Interface ( SPI ) mode and Secure Digital ( SD ) mode, we used the SD mode which is more compatible for media storage. We need 11 wire connections including a ground and a VCC between the SD module and the TI evaluation kit. The SD card needs a supply vol tage of 2.0 to 3.6V from the board. CC1110 handles incoming RF data packages while controlling the data saving operation. Since the RAM of CC1110 is too small (only32kB) it is not sufficient to hold the whole image frame. Our 640 by 480 pixel picture size varies from 55kB up to 170kB. So we have to perform online storage. The SD card needs to be formatted into FAT32 file system before being loaded into the module. The software is modified from a public ly available data transmission sample code on TI websi te. SD card operation sections are added into the code.

PAGE 117

117 A B Figure 5 1 2.4 GHz wrapped patch antenna with inductive loading A ) Sch ematic B ) Fabricated device: Dimensions: a = 18.5mm; b = 12.5mm; c = 5mm; d = 1mm; e = 2mm; the cylinder diameter is 10mm in B ) (the ground plan e is not shown for clarity in A )) Figure 5 2 Measured return loss A B Figure 5 3 Simulated antenna radiation pattern A) H plane. B) E Plane. -20.00 -15.00 -10.00 -5.00 0.00 1.00 2.00 3.00 4.00 5.00 Flat Rolled,D=5mm Frequency / GHz Return Loss/ dB

PAGE 118

118 A B Figure 5 4 Proposed wrappable capsule endoscope system A ) Bare LCP based substrate with the wrappable patch antenna B ) S ystem with electronics soldered. A Pressure Sensor ADC:AD7746 Transceiver:CC2510 Wrappable Antenna

PAGE 119

119 B Figure 5 5 Receiver side solution A ) PC side software showing sensor status and pressure sensor calibration function B ) Receiver USB controller A B C Diode Power Conditioner Camera With LED Camera With LED CC1111

PAGE 120

120 Figure 5 5 Wireless endoscope based on hybrid circuits. A) Rigid circuit parts. B) Wrappable antenna. C) Assembled system A B Figure 5 6 Temperature characterization of the capsule endoscope A) Temperature measurement. B) Pressure measurement 20 25 30 35 40 45 0 5 10 15 20 Wireless Sensor Thermometer Time / Minute 1.71 1.72 1.73 1.74 1.75 1.76 1.77 0 10 20 30 40 50 Pressure in Atmosphere / mmHg Capacitance / pF

PAGE 121

121 A B Battery Power Management CC1110/CC2510 Processor and Radio Camera Pressure Sensor

PAGE 122

122 C D

PAGE 123

123 E Figure 5 7 Capsule endoscope block diagram A) System function diagram. B) Schematic of CC1110 SoC. C) Pressure sensor c onnection. D) Power conditioning module. E) Camera module. A B

PAGE 124

124 C D Figure 5 8 Camera imaging distance optimization: A) Image taken at 5 cm away, B) 2.45 cm away, C) 1 cm away, D) 1 cm away with optimized software

PAGE 125

125 Figure 5 9 Wireless data collector for the wireless endoscope

PAGE 126

126 CHAPTER 6 SMART MOUTH GUARD 6.1 System Requirements Intraoral sensing is of great interest for dentistry pathologies. For bruxism detection and management, biting pressure is of significance. Bruxism involves the activities of grinding or clenching the teeth especially when people are sleeping, which could be considered as a sleep disorder as well. Bruxism is regarded as a reason for temporomandibular disorder (TMD) and associated with chronic jaw pain and headache. As the second most common musculoskeletal pain, TMD affects abou t 10% of the general populati on [4 9 ] It should be noticed that people are also grinding teeth subconsciously during the day time whic h is also considered as bruxism [ 50 ] Wired intraoral pressure sensing sys tems have been reported before [ 51 52 ] However, a wired system confines the patients from moving freely and is not preferable for measurement during sleep And also, due to the presence of a connection wire the measured data are subject to doubt. A p recise and wireless sensing system for the biting pressure is the kernel part fo r a n mHealth system targeting to monitor bruxism. BiteStrip is a commercially available jaw muscle electromyography (EMG) which could stick to the during sleep, and the system would record the muscle movement during sle ep [5 3 ] However, the accuracy of this strip is limited by the electrode positio n, posture and skin resistance [5 4 ] while it only tells if teeth grinding occurs, and no detail information such as the biting pressure, time and exact frequency is available F urther, BiteStrip does not provide any protection function for bruxism and not offer wireless data recording functionality either. Thus an intraoral sensor system integrated in a mouth guard or splint is preferred which could deliver the best comfort as well as protection t o bruxism. An alternative method is to use a piezoelect ric sensor for bruxism sensing [5 5 5 9 ] However, the piezoelectric sensors are not suitable for continuous pressure detection due to its

PAGE 127

127 abrupt nature and those sensors are not suitable for certain pat ients such as children and senior citizens whose biting force might be too weak to excite the piezoelectric sensors. Recently, sensor systems based on force resistance transducers have been introduced for bruxism monitoring [5 4 60 61 ] However, it can be seen that this kind of transducer has poor linearity, after a certain force range (about 40 N) the output from the transducer tends to saturate, which is not sufficient to measure a bruxism pheno menon [6 2 ] F urther those force to resistance transducers a re very complicated to implement and only a small amount of sensor s can be fitted into a single splint, and maximally two sensors of such kind s ha ve been reported [5 4 61 ] A number of patents are also filed for bruxism management M ost of those patents fall into one of the above categories. From the practice point of view, most of the released wireless bruxism monitoring systems do not have power management functions, which prevent them from performing long term measurements. To our knowledge, no measure d data for over 10 minutes has been reported yet out from a wireless system. Also all the proposed systems are based on traditional rigid PCB s which cannot easily and wearing comfort cannot be guaranteed F urther, those r igid implants may cause unwanted injury or trauma to the wearer. Meanwhile, all the wireless bruxism monitors reported to date have either piezoelectric sensors which are not suitable for bruxism monitoring due to its abrupt nature, or complicated and expe nsive transducer designs, and all the systems can only adopt the very limited number of sensors T he sensor accuracy and operation range, to our knowledge, are not characterized as a function of the human biting force. None of the current existing systems has a mature networking consideration N o data encryption and network authentication is considered in most works In such case, undesired parties can obtain the measured private health information data easily

PAGE 128

128 M eanwhile without proper network protocol sup port, it is very difficult to achieve a wireless system with low power consumption. This work proposes a low cost and easy operat able mHealth system for bruxism self monitoring and diagnosis. It is composed of a USB dongle which could be plugged into a mo bile terminal such as a laptop computer or integrated into a cell phone sleeve; a spli n t integrated with wireless sensing systems based on the flexible PCB; a user friendly interface for data collection and system control. A c apacitance based sensor array with a low cost design is used where those sensors are simply metal pads with silicone coating. Sensor performance is characterized G ood linearity and sensitivity is achieved. Flexible circuit design is used to best fit to the mouth shape F lexi ble RF circuits including transmission lines and antenna s are designed and characterized. Proper power management units are used for long time measurement capability. SimpliciTI from Texas Instruments (TI) has been adopted as the network protocol which offers flexible network topology as well as network authenti cation It is a preferred protocol for a small scale low power network. 6.2 Wireless Mouth Guard Based on Capacitive Force Sensor 6.2.1 Sensor Principle There are two kinds of capacitors that could be adopted for sensing purposes, namely, a fringing field capacitor and a parallel plate capacitor as shown in Figure 6 1. The principle for a capacitive pressure sensor is shown in Figure 6 1 A W hen pressure is applied on to a grounded object (could be human body or other large object), it travels toward the sensor, and part of the fringing field will be grounded T hus the capacitance drops when the grounded object is getting closer to the sensor which indicates an increased pressure level F or a parallel plate capacitance sensor, the external pressure alters the distance between two plates C apacitance increases when the higher pressure is applied.

PAGE 129

129 It can be seen that the fringing capacitance sensor requires a sophisticated sensor design I t has been demonstrated in [6 2 ] that the gap between electrodes, the elect rode area and the shape are all contribut ing to the sensor performance, and thus complicated sensor shape design and careful manufacture procedure are needed for such a sensor. Traditional parallel plate capacitor based sensors as reported in [6 3 ] are eas y to fabricate but show poor linearity [6 2 ]. In this work, we use a modified version of a parallel plate based capacitance sensor for the intraoral pressure sensing purpose. The proposed pressure sensor is shown in Figure 6 2 B For a traditional pressure sensor, two electrodes are connected to the opposite polarity of a voltage source and the electric field is formed between the plates as shown in Figure 6 2 A Since none of the electrode s really touches the earth, we call this configur ation a floating parallel plate capacitor while in our proposed version, the sensor is simply a metal pad covered by a certain elastic material as shown in Figure 6 2 B In this capacitor, the human body, which is also conductive, serves as the other electr ode, and we call this configuration a non floating parallel plate capacitor. The physic al deformation and the corresponding force can be correlated as Equation (6 1). ( 6 1 ) where F is the applied force, is the original linear length of the object along the force application direction is the length variation along the force application direction is the cross section area where the force is applied and Y is the elastic modulus of the elastic materi al Equation ( 6 1 ) reveals a linear relationship between the length variation and the applied force. The c apacitance is defined as the ratio between the amount of charge enclosed in a certain area and the voltage devel oped cross it. ( 6 2 )

PAGE 130

130 Due to the separation by the elastic dielectric material, once the voltage V is applied on the sensing pad, the electric field E will be formed between the sensing pad and the human te eth. Since the potential inside the conductor pad must be zero to satisfy the boundary condition, we only have surface charge accumulation on the pad, and the surface charge density denoted as is described as Equation (6 3) ( 6 3 ) where q is the total charge on the pad surface in Coulomb and A is the surface area where biting force would be applied to, in meter square Since ( 6 4 ) W e now need to find out the electric field distribution due to the presence of the metal pad. From the ( 6 5 ) where D is the the electric flux density in newton per volt meter, and has a relationship of D = where is the electrical permittivity of the elastic material. From Equation ( 6 5 ), we have ( 6 6 ) where is the distance from a point on the metal pad and a point in the elastic material. Substitute Equation ( 6 3 ) and ( 6 5 ) into Equation ( 6 6 ): ( 6 7 ) From Equation ( 6 4 ), the voltage developed in the elastic material is: ( 6 8 ) It is surprisingly to see here in Equation ( 6 8 ) that the voltage between a point in the elastic material and the ground (human body) does not depend o n distance but it seems that the

PAGE 131

131 elastic material has equal potential everywhere with respect to the human body T his is a very important observation T his feature allows a linear relation between the capacitance change and the thickness variation. Substitute Equation ( 6 8 ) back into Equation ( 6 3 ), we have: ( 6 9 ) Equation ( 6 9 ) gives a capacitance equation that is identical to the traditional floating capacitor. Equation ( 6 2 ) and ( 6 9 ) suggest a linear relationship between the biting force applied on the elastic material and the capacitance variation. 6.2.2 Sensor Fabrication From the discussion above, the newly proposed biting force sensor is extremely simple and easy to fabricate. It is simply a metallic pad with certain biocompatible elastic material coating. Such a configuration is also compatible with the standard industry PCB fabrication procedure and is suitable for large scale low cost manufactur ing For prototyping purposes, sensing pads with different dimensions are fabricated on a flexible Rogers RO3003 substrate (10 mil s thick) as we have discussed before. The fabric ation is done using a milling machine S100 from LPKF Inc. Since the substrate is very thin, extreme care should be taken while adjusting the drill bit depth for the milling machine. The system schematic and layout is first designed in Altium Designer. We t hen export necessary fabrication files. This includes several layers of gerber file (*.ger) which are used to describe the circuit layout on each layer and a NC dill file (*.cam) which includes the via information. Those files are imported to CircuitCam to calculate the milling path, a *.LMD file is generated containing the milling path information. Finally, we use CircuitPro to read in the LMD file and control the milling machine to mill. The fabricated sensors are as shown in Figure 6 3

PAGE 132

132 There are five se nsor pads with pad sizes of 3mm by 3mm 4mm by 4mm 5mm by 5mm 6mm by 6mm and 7mm by 7mm as shown in Figure 6 3 S ince the substrate is flexible, we may bend this sensor pad array freely as shown in Figure 6 3 B T his flexibility delivers best wearing com fort, while it makes the sensors stick to the spli n t well to generate accurate measurement results. It should be noticed that the back side copper of this substrate is not removed. We will keep this back side copper for noise shielding purposes. It can al so be seen that a transparent, flexible and elastic material has been coated on the sensor pad array. The material here we used is polydimethylsiloxane (PDMS) which is also known as silicone. It is a kind of organosilicon compound that is bio compatible. T he PDMS we used here is Sylgard 184 from Dow Corning Inc. which has a relative permittivity of 2.65. 870 kPa The PDMS thickness varies from 3 mm to 4.5 mm for full characterization. The sensor array characterization is done on a digital scale JSHIP 332us ( Jennings Inc.) Different levels of force are applied on each sensor and the output from the ADC is recorded. As we have discussed from the previous section, the force applied has a linear relationship with the output capac itance H ence here we do not need to convert the ADC output into the capacitance value I nstead, we just record the normalized output and then we can setup a mapping between the applied force and the ADC output value. The characterized sensor behavior is s hown in Figure 6 4 6.2.3 Sensor Electronics The sensing circuit is composed of mainly a transceiver, a microcontroller, multiple ADC s or an ADC with multiple input channels, passive RF components such as a high performance antenna, transmission line, balun and so forth. The system diagram is shown in Figure 6 5

PAGE 133

133 The system is implemented on a flexible substrate Rogers RO3003 which has a dielectric constant of 3 and a thickness of 10 mil s (0.25mm) T he subst rate is flexible and can be reconfigured to fit the mouthguard profiles for different patients. The circuit is fabricated using a S100 milling machine from LPKF Inc. I t should be noticed that extra care should be taken when mill ing on a flexible substrate A new vacuum table needs to be used to mill such a board and careful drill bits depth adjustment is a must. To have a smaller system size, discrete components with small packaging (SMT 0201) are used which corresponds to a physical dimension of 250 500 Figure 6 5 The whole system is shown in Figure 6 6 below. It should be noticed that the system is in the deep sleep mode in most of the time, when several microamperes of current is drawn. A magnetic reed switch is used to wake up the system from the sleep to active mode. The main reason is that once the system is integrated into a mouth guard, it is difficult to have a physical switch A nd in this system a contact less remote switch is used. The system is then molded into a mouth guard in PDMS material s the mouth guard integrated with sensors is shown in Figure 6 8 6.3 System Characterization The mouth guard is first worn by the author and the inhabita nt bite s for 4 times Data acquisition is done on a computer. B iting was successfully recorded as shown in Figure 6 9. Then the inhabitant is asked to clench the left side, the right side and the whole teeth Sensor 1, Sensor 2, and Sensor 3 are located at the left side, the front, and the right side of the mouth guard, respectively. T he measured results are shown in Figure 6 10 A s shown in Figure 6 10, while clenching on the left side, S ensor1 (on the left) and S ensor 2(front) have prominent res ponse while clenching on the right side, S ensor 3 (on the right) and S ensor 2(front) have prominent response. When clenching on the whole mouth guard, all

PAGE 134

134 sensors have prominent response This experiment indicates that there is very few coupling between differ ent sensors and teeth clenching locations can be identified with the proposed system. 6.4 Blue Guard: A Bluetooth Enabled mHealth System for Bruxism Management With the development of smartphones, it is feasible to perform Bruxism at home by a cell phone The mobile phone based health monitoring system is also known In such a system, the smartphone works as a data hub that collects and processes the sensor data, generate health condition reports. In this work, we use the concept o f mHealth system to manage the Bruxism. The system block diagram is as shown in Figure 6 11. The Bluetooth enabled Bruxism sensor and transceiver system are integrated in a mouthguard, the sensor system captures Bruxism activity and records the biting forc e. This force is digitized and transmitted to the smartphone. The smartphone used here is based on Android Jellybeans build. The sensor data are summarized and could be transmitted to the designated health care provider on demand. In Bruxism detection, sig nal detection in an oral environment experiences common mode noise due to the movement of oral tissues. Therefore, a reliable noise cancelation method is a n important specification in a high performance Bruxism management system. Blue Guard employs a RC discharge method for Bruxism monitoring. The system accommodates a low power microprocessor ( MSP430BT5190 TI Inc.) where data digitization and conditioning is realized by software and no hardware analog to digital data converter is needed. A quasi differ ential input scheme is used for low noise performance. The e lectrically conductive body can be utilized for the implementation of a parallel plate based capacitive sensor a s shown in Figure 6 12 i.e. the human tissue of the mouth serves as an electrode of the sensing capacitor and only one conductive pad is sufficient. Nontoxic

PAGE 135

135 biocompatible PDMS is coated on the surface of the metal pad. PDMS is elastic and the dielectric constant is 2.9. With a pad size of 5 mm by 5 mm and a PDMS thickness of 1m m, theore tically the initial capacitance without biting is about 3pF. T o improve the noise performance, a ground plane is placed beneath and around the sensor pad which results in an initial capacitance of approximately 10 pF due to parasitic capacitance A resistor is used to form a single time constant RC network as shown in Figure 6 1 3 T he capacitance change is detected by measuring the time constant of the RC network using a low power microprocessor Port 1 is a general purpose I O port and can be reconfi gured as either input or output state. The voltage cross the capacitor VC is shown in Figure 6 13 B ( upper ) In the first stage, the general purpose IO pin on the microprocessor is configured to be an output high, and +3.3V DC voltage is applied on the capa citor S ince the resistor is connected in parallel with the capacitor, the capacitor is charged to + 3.3V quickly. In the second stage, the IO pin is configured to be an input, and the capacitor discharges through the resistor. By measuring the charge and discharge time which is proportional to the time constant = RC, the capacitance value is measured. Here, t he discharge time is measured by the microprocessor. As shown in Figure 6 1 3 B the clock frequency of the counting timer is chosen to be 16 MHz, and the clock period is 0.0625 s. With a 6 M the time constant of the RC network is 60 s. This large resist ance sets a proper time constant and also limits the discharge current I n the 6 M case, the discharging current will be 5 5 0 nA which greatly reduces the overall power consumption while having great accuracy For counting the discharging time, the Subsystem Master Clock with a frequency of 16MHz is used and T imer A is setup for discharging time measurement.

PAGE 136

136 The more detail mechanism is following. The a ctive pin is first configured as an output high to charge the capacitor. After full y charging, a snap shot of Timer A is taken, and a falling edge triggering interrupt is enabled on the IO pin where it is set to an input, and the capacitor discharges through the resistor After the capacitor voltage falls below a certain value, a falling edge is detected, and another snapshot of Timer A is taken. The difference between those two snapshots is the number of system clock periods used during di scharging which is proportional to the discharging time. The greater biting force, the larger capacitance, the longer discharge time and the more clock counting output become It should be noticed that although the small discharging current help s the reduc tion of power consumption, a more stringent requirement is p laced on the microprocessor, especially a processor with low leakage current on IO pin and accurate timer. MSP430BT9150 has a leakage current on the IO pin as low as 50 nA which is much lower than the discharging current of 550 nA, and thus the uncertainty due to pin leakage current is very small [ 61] A major concern in this method lies in the common mode noise appearing on the sensor pad. To cancel out the noise, a pair of pad s is used instead of a single pad. As shown in Figure 6 14 in the quasi differential input mode, one pad works as the sensor pad while the other pad is connected to a neighbor IO pin to be ground. In this case, a time constant corresponding to Active pad 1 i s measured first. T hen the active and partner pad s exchange the role, and another time constant 2 is measured. The final result is obtained by averaging those two results and thus common mode noise present on both pads is canceled. Power consumption is also optimized. Since the discharging period takes most of the machine time, we can put the microprocessor in a low power sleep mode when measuring the

PAGE 137

137 discharging time during which the microprocessor consumes only approximately 69 A, while in an active conversion mode it consumes approximately 1.68 mA. A Bluetooth module PAN1325 is integrated with the capacitive sensor system and a wireless mouth guard is implemented. A s show in Figure 6 15 A the circuit has an overall dimension of 15 mm by 20 mm and has been embedded into a mouth guard for Bruxism monitoring. The microprocessor is MSP430BT5190, which is designed for Bluetooth and Bluetooth Low Energy stack application. It has 100 pins. In Figure 6 15 B, bus translator based on TI SN 74AVCB164245 is used. Figure 6 15 C shows the configuration of the Bluetooth Module. The programming of the hosting software running in the Android cell phone is done in Eclipse which is a open source IDE for Android platform. The Blue Guard application is as shown in Figure 6 16. The application can be launched by simply tap on the icon. The initial window after application launching is as shown in Figure 6 16 B. First action is to turn on the tton, this sends a command to the Blue Guard sensor and the sensor begins to collect data and transmit the sensor data. Next phone will search for all pair abl e Bluetooth device near by, tap on the Blue Guard hardware ID to connect. In this case, the application will require the user to input a password, once password matches, the Bluetooth connection is established between the cell phone and the sensor, real ti me sensor data are plotted as shown in Figure 6 16C. To characterize the sensor, standard force levels are applied on the sensor surface and the sensor output is recorded. To characterize the repeatability and reproducibility of the sensor,

PAGE 138

138 each force level is applied for 10 times and the statistical data are a s shown in Figure 6 17 It can be seen that 1) The sensor shows good linearity. 2) When the force applied to the sensor gets greater, the measurement uncertainty increases, the standa rd derivation of the sensor output increases from 0.12 (5N) to 0.20(50N), this observation has never been reported before. In this work a wireless mouth guard for Bruxism management is demonstrated A sing le time constant RC network is used for biting forc e monitoring. Power consumption is reduced by using larger discharging resistance thus achieving a Nano ampere level di s charging current T he microprocessor is held in a low power mode while measuring the time constant for the further suppression of power consumption Using quasi differential mode sensor pairs reduces measurement noise No additi o nal hardware is needed to implement this sensor, which aids the reduction of system size and power consumption. 6.5 A Wireless and Battery Free Mouth Guard for Bruxism Management Based on Flexible Metamaterial Particles Battery powered systems are usually limited by the battery capacity I n this section, a battery free Bruxism mouth guard is designed, fabricated and characterized. Split Ring Resonator (SRR) is one of t he metamaterial particles that would offer negative permeability. T he narrow bandwidth feature of an SRR makes it ideal for wireless sensing due to its high selectivity The SRR is a sub wavelength structure and it is often used for device size reduction. The SRR can be tunable when it is loaded with a tunable component such as a varactor All those attributes make the SRR as a good candidate for a compact high performance wireless sensor.

PAGE 139

139 In this work, an SRR based compact wireless and battery free sensor is integrated into a mouth guard for Bruxism management. The integrated mouth guard reduces damage s to the teeth muscle, and jaw bone due to excessive grinding and clenching, and monitors the biting activities and behavior wirelessly without any battery integrated The sensing data can be read out by a receiving antenna and a frequency response measurement instrument. An SRR is composed of two concentric rings with a slit caved on each of rings The outer ring has another slit, as sho wn in Figure 6 18 A where a varactor is loaded to achieve a tun ing fu n ction As shown in Figure 6 18 B the SRR with a varactor can be modeled as an L C tank with a tunable element Cv which represents the varactor capacitance. Ls and Cs represent the inherent inductive and capacitive components of the SRR and the resonance frequency of an SRR can be determined by Equation ( 6 10 ) [62] : ( 6 10 ) where C tot represent s total e ffective capacitance. When reversely biased, a diode demonstrates a tunable junction capacitance and it can be determined by: ( 6 11 ) where N A and N D are the doping densities of n type and p type semiconductor s respectively, Va is the negative biasing voltage, i is the buil t in potential, s is the electrical permittivity of the semiconductor Equation ( 6 11 ) indicates that the junction capacitance is reversely relate d to the biasing voltage. With varactor loading, the resonance frequency of the SRR can be modulated by the biasing voltage.

PAGE 140

140 P i ezoelectric refers to the charges accumulated in materials due to mechanical stress or deformation of the material. In order to bias the varactor, we use a piezoelectric sensor film ( MEAS DT Measurement Specialties, Inc.) This is a piezo film for dynamic strain and vibration detection. The piezoelectric film is characterized first. We use standard weights to mimic the human biting force. As different levels of weights are applied, the piezoelectric sensor output voltage level varies nearly linear ly as shown in Figure 6 19 It should be noticed that the film has been coated with a layer of PDMS with a thickness of 2mm before characterization. PDMS is an elastic and biocompatible material and makes the sensor less sensitive to ambient noise. I t also prevents the sensor from direct contact with the intraoral biochemical media The varactor ( SMV1232 Skyworks Inc ) is a kind of hyper abrupt junction tuning diode with low series resistance and is ideal for high Q resonator tuning. As shown in Figure 6 19 the piezoelectric film has a nearly linear output voltage up to 36 Kg ( 70 lb) applied weight, where 3.4V DC out put voltage is observed across the output terminal of the film. This range covers the human biting force range from 10 lb. to 100 lb[53]. From Figure 6 20 it can be seen that the varactor demonstrates large capacitance variation up to a reverse biasing vo ltage of 3.4V, which contributes to good sensitivity in the operating force range The SRR loaded with a varactor is integrated with a dipole antenna as shown in Figure 6 21 A The dipole antenna is used to enhance the coupling between the sensor and the external applicator antenna. The system is fabricated on a high frequency laminate (RO 3010, Rogers Inc.), which is based on ceramic filled Polytetrafluoroethylene (PTFE) with a dielectric constant of 10.2, a loss tangent of 0.002, and a thickness of 10 mil s (0.25 mm). Due to the small thickness,

PAGE 141

141 the sensor system is flexible and can be bent to fit the mouth guard curvature as shown in Figure 6 21 B The SRR is designed to reso nate at a 5.8 GHz band, the outer diameter of the SRR is 6 mm (0.31 ), and the SRR trace width is 0.3 mm. Then the whole mouth guard consisting of a piezoelectric film, a tunable SRR and a dipole antenna is assembled as shown in Figure 6 22 M e asurements are performed using an Agilent E5071C VNA A broadband horn antenna is used as an applicator antenna, and S11 is measured. The a pplicator antenna is kept 30 mm away from the wireless mouth guard. Prior to measurements, the VNA is calibrated and t he measurement setup is shown in Figure 6 23 Figure 6 24 shows the return loss spectra as a function of the different force input. The force applied on the piezoelectric film modulate s the output voltage of the film, which changes the equivalent capacita nce of the varactor according to Eq uation ( 6 11 ), resulting in the resonance frequency shift. The relationship between the weight applied on the piezoelectric sensor and the output is shown in Figure 6 25 A nearly linear performance is obtained The SRR tunability is limited by its nonlinear behavior, and the resonance frequency stays nearly constant after a certain capacitance value [ 6 3 ] This phenomen on would mainly affect the small force detection T he sensitivity of the system is limited by the nonlinearity of the tunable SRR. In this work, a compact wireless and battery free mouth guard for Bruxism management is presented based on a flexible metamaterial particle integrated with a varactor and a piezoelectric transducer A linear size redu ction up to 38% is achieved comparing to a half wavelength resonator. Good linearity is observed from the sensor, and this system offers a platform for wireless battery free biomedical sensing for the force and strain.

PAGE 142

142 6.6 Future Work T he system is based on SimpliciTI which is a low power and simple protocol H owever, there is a demand to have home monitoring and self diagnosis system for a patient to manage Bruxism at home. The system is being considered to be implemented based on the Bluetooth protocol. W it h Bluetooth, more current draw is expected. One thing we need to consider is to smartly design the current drawing profile W e may leave the system in sleep for more time. However, the drawback of this method is to have a less record time resolution, and a certain amount of Brux will be lost. If we review Figure 6 5 A we may setup an interruption once the detected capacitance exceed s a certain threshold, or once huge capacitance variation takes place This means that when the patient teeth are touching dow n to the sensor pad, the system can be activated by this interrupt, and the pressure measurement circuit starts to work and the sensing data are recorded

PAGE 143

143 A B Figure 6 1 Ca pacitance Sensors A ) Fringing B ) Parallel plate A B Figure 6 2 Different kinds of parallel capacitive sensors A ) floating B ) non floating A + Excitation Detection Elastic material

PAGE 144

144 B Figure 6 3 Fabricated flexible biting force sensors A ) top view B ) side view after bending Figure 6 4 Capacitive sensor characterization, PDMS thickness is 3 mm A 0.8 0.85 0.9 0.95 1 1.05 0 20 40 60 80 100 120 3mm x 3mm 4mm x 4mm 5mm x 5mm 6mm x 6mm 7mm x 7mm Force / Newton Normalized ADC Output 5 Channel ADC: AD7148 2.4GHz SoC CC2510; INT I 2 C Antenna

PAGE 145

145 B C Figure 6 5 System block diagram A ) In mouth sensor part B ) Wireless USB type controller and data collector part C ) Schematic in Altium designer A B Figure 6 6 RF Balun on the system A ) Before soldering B ) After soldering

PAGE 146

146 A B C

PAGE 147

147 Figure 6 7 Fabricate d system A ) Front B ) Back C ) System activation by a magnet A B Figure 6 8 Mouth guard with circuitry integrated A ) Front B ) Back Figure 6 9 Biting characterization -1.5 -1 -0.5 0 0.5 1 1.5 182790 182792 182794 182796 182798 182800 182802 182804 182806 182808 182810 Sensor 1 Sensor 2 Sensor 3 Sensor 4 Sensor 5 Time: HHMMSS Normalized ADC Output

PAGE 148

148 Figure 6 10 Teeth clenching at different position s Figure 6 11 Block diagram of the Bruxism management mHealth system Figure 6 12 C apacitive sensor for Bruxism monitoring 0 0.2 0.4 0.6 0.8 1 1.2 0 2 4 6 8 10 12 Sensor 1 Sensor 2 Sensor 3 Time / Second Normalized ADC output + Elastic material ( PDMS ) Clock Vc/V time/s time/s

PAGE 149

149 A B Figure 6 13 RC network for Bruxism monitoring A) Schematic. B) Clock diagram. Figure 6 14 Quasi differential input A

PAGE 150

150 B C D E Figure 6 15 Wireless Bruxism management system A) Processor. B) Bus translator. C) Bluetooth module. D) Application LED and Reset switch. E) Assembled Blue Guard sensor BT module MSP430 processor Biting force sensing pad

PAGE 151

151 A B

PAGE 152

152 C Figure 6 16 Android application for Bruxism management: A) App logo, B) Device Pairing, C) Sensor data real time plot

PAGE 153

153 Figure 6 17 Measured output as a function of applied force on sensor s error bars show the repeatability A B Figure 6 18 Varactor loaded split ring resonator A ) Schematic B ) Equivalent circuit model Figure 6 19 Piezoelectric f ilm c haracteri stics 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 5 25 45 65 85 105 Force Applied[N] Normalized sensor output 0 1 2 3 4 0 10 20 30 40 Applied Weight [Kg] Output Voltage [V] + V a Ls C s Cv

PAGE 154

154 Fi gure 6 20 Varactor t unability A B Figure 6 21 Varactor l oaded t unable SRR A ) F a bricated s ystem o n a p lanar substrate B ) Bended one showing its flexibility Figure 6 22 Assembled wireless battery free m outh g uard Figure 6 23 Measurement s etup 0 1 2 3 4 5 0 2 4 6 8 10 12 14 Reverse Biasing Voltage [V] Junction Capacitance [pF] Varactor Broadband Applicator Antenna Mouth Guard with Sensor

PAGE 155

155 Figure 6 24 Measured f requency r esponse at d ifferent w eight l evels Figure 6 25 Resonance frequency shift as a function of the applied weight -45 -35 -25 -15 -5 5.6 5.8 6 6.2 6.4 36Kg 0Kg Frequency [GHz] LOGMAG(S11) 6.1 6.12 6.14 6.16 6.18 6.2 6.22 6.24 0 10 20 30 40 Weight Applied [Kg] Resonance Frequency [GHz]

PAGE 156

156 CHAPTER 7 CONCLUSION Different kinds of b iomedical telemetry systems are explored in this work. Modern medical devices need to have good performance while maintaining their small overall size. These stringent requirements give rise to difficulty in the system design. Great research efforts h ave been exerted to implement systems with compact size yet high performance. This includes high measurement accuracy, novel sensing methods and extended device operation time without battery replacement Different platforms are explored including a whol e solution for Bruxism and oral health condition monitoring system and a low cost high performance wireless endoscope. This work gives a comprehensive understanding and research on the components, subsystems and overall system of biomedical telemetry sys tems. The first two chapters build certain foundation for this research. Due to the dispersive and dissipative nature of the human body, an accurate human body model for high frequency circuit s is studied and implemented. Through this work, we use the huma n body numerical model from Ansys. Th e model s are available for different sub versions with different levels of accuracy. As we have stated above, the stringent requirements for biomedical telemetry system s make them very difficult to design and implement. W e use metamaterial particles to overcome those technical difficulties such as size reduction, and high sensitivity sensors etc. Metamaterial refers to structures which have properties that cannot be readily obtained from nature. In the microwave society metamaterial particles refer to sub wavelength structures with near zero or negative permittivity and / or permeability. Since they are sub wavelength structures, metamaterial particles are ideal for microwave circuit miniaturization. This feature meets the requirement of small device size. Metamaterial particles such as SRR and CSRR are of resonator nature and thus can be modeled

PAGE 157

157 as LCR tanks. If certain tunable components such as a varactor are loaded on such a particle, those particles serve as good el ectrically small sensors. Due to the high Q nature, those resonators are very sensitive as sensor units. To demonstrate the tunability, we implemented two arrays of SRRs sta cked back to back in a rectangular waveguide. By changing the relative position of the arrays, mutural coupling is also changed. Thus the transmission at the resonance frequency is controllable. We then demonstrated an electrically small and tunable antenna based on SRR particles. A varactor is loaded on a SRR for tuning purposes. A tun able SRR antenna is investigated both theoretically and experimentally. The proposed antenna has a variable radiation frequency from 1.81 GHz to 2.14 GHz with a tuning voltage between 0 and 9.5 V. Simulated data show good agreement with measurement ones. Besides tuability, the proposed antenna has achieved a linear dimension reduction by 59.3 % compared with the monopole counterpart. While Chapter 2 focuses on passive metamaterial particles, Chap ter 3 investigates active components. Since metamaterials are passive device s external stimuli are a must to excite those particles. A broadband amplifier is often used for such a purpose. However, due to the un avoidable packaging parrasitics and other n on ideal factors, major broadband amplifiers show a certain gain roll off at high frequency band that is, the amplifier gain is high at low frequenc ies but reduces gradually at high frequenc ies We designed a low noise and compact frequency equalizer netw ork to compensate this gain roll off. With the area of 6 mm 8 mm, a gain flatness of 1.1 dB is demonstrated from 100 MHz to 7 GHz with a nominal gain of 33.5 dB. The proposed frequency equalizer has a simple architecture for design and implementation b ut performs very well.

PAGE 158

158 The a ntenna is always difficult to implement in a biomedical telemetry system. On one hand the radio frequency behavior of antennas could be very different in the human body than in free space O n the other hand, since the orientation of the device is not predictable for many of the biomedical sensing system cases, blind spot s of observation need to be minimized. In addition, medical instruments need to meet stringent FCC and FDA regulations. One of the most important figure of merit is the electromagnetic compatibility of the device. That is to say the device cannot interfere other surrounding systems and it should not be susceptible to external interference. Certain EMI isolation is a must in a medical telemetry system. In Chapter 4, we begin with a patch antenna and smartly fold it into a rectangular waveguide shape. By doing so, a n EMI shielded cavity is formed and all the circuits loaded into this antenna cavity are protected by the ground plane of the antenna inherently. Meanwhile, this antenna produces a quasi omnidirecti o nal pattern which assists blind spot miniaturization. Then several different approaches including inductive l oading, via loading and metamaterial particle loading methods are used for antenna miniaturization. Based on the antenna designed in Chapter 4, we have further developed a new concept: wrappable electronics. The idea is to take full advantage of flexible circuits. Instead of designing circuits onto different layers and then stack up, we design all the electronics on one piece of flexible circuit board and the final assembly is done by simply rolling or folding the circuit s into the desired shape. A wrapp able wireless endoscope with multiple sensors and a camera is demonstrated in Chapter 5 Comparing to its traditional counterpart, the wrappable version has a simpler design, and better performance. Chapter 6 focuses on another family of telemetry system: wireless oral health monitors. We have first demonstrated a wireless capacitive system for Bruxism monitoring. Then, we

PAGE 159

159 further developed a new generation of smart mouth guard system based on RC time constant method. Finally, we use a piezoelectric membra ne and metamaterial particles as sensors, and a wireless and battery free version of smart mouth guard is demonstrated. The contribution of this work includes: (1) Controllable transmission in waveguide based on metamaterials with in situ tuning has been successfully demonstrated. ( 2 ) We are the first group demonstrating a wrapped patch antenna with excellent performance for wireless endoscope applications B ased on this antenna, the concept of wrappable electronics is proposed. ( 3 ) We first demonstrated the time constant method for biting force monitoring. ( 4 ) We are the first group demonstrating the battery free and wireless biting force monitoring system based on a piezoelectric membrane and metamaterial particles. In the future, energy would be one of the most important issues that a medical device needs to face. Antenna s with multiple functions including wireless communication and wireless power reception will be a research trend. Although we have some preliminary explor ation in Chapter 4, this antenna still need s whole wireless power conditioning circuitry to achieve a high performance battery free wireless endoscope. This may need a high performance IC and a high performance super capacitor that can stand very qui ck charge and discharge. We have demonstrated that human biting activity can be monitored by a cell phone via the Bluetooth communication protocol. However, in the future, reader circuitry can be integrated in to a cellphone sleeve. This system can activat e a wireless sensor and measure the response as desired. In this case, a wireless battery free mHealth system can be implemented. All those systems will greatly simplify the in home diagnosis and disease management procedure, and life quality of certain pa tients will be greatly improved.

PAGE 160

160 LIST OF REFERENCES [1] Naicker S, Plange developing countries Africa 64. [2] Healthcare Sensor Networks : Challengers toward practical implementation CRC Press, 2012. cellular telephone subscriptions 2010 2011 http://www.itu.int/ITU D/ict/statistics/ Mobile Health Te chnologies http://obssr.od.nih.gov/scientific_areas /methodology/mhealth/index.aspx [6] Evaluating Compliance with FCC Guidel ines for Human Exposure to Radiofrequency Electromagnetic Fields available online: http://transition.fcc.gov/Bureaus/Engineering_Technology/Documents/bulletins/oet65/oet 65b.pdf [7] Vorst, A. vander, et al., RF/Microwave interaction with biological tiss ues John Wiley & Sons, 2006. [8] John Wiley & Sons, 1989. [ 9] Ryckaert, J., De Doncker, P., Meys, R., de Le Hoye, A., Donnay, S. "Channel model for wireless communication around human body," Electro nics Letters vol.40, no.9, pp. 543 544, 29 April 2004 London : Springer, 2008. plastic sub Scientific Reports, vol.2, 2012, p.622. crystal silicon for high performance el Science vol. 311 pp.208 212, 2006. [13] Sun Y, Choi nanoribbon Nat Nanotechnol vol.1, pp.201 207,2006. effect tr ansistor r ealized by Science vol. 265 pp.1684 1686, 1994. Appl Phys Lett vol. 24, pp.610, 2002

PAGE 161

161 [16] Drury cost all polymer integrated circuits Appl Phys Lett vol. 73, pp.108 110, 1998. [17] William D. Callister, David G. Rethwisch ,"Materials Science and Engineering An Introduction," John Wiley & Sons, 2010. Solid State and Materials Science, vol.6, issue 6, December 2002, pp. 545 551. available online: http://www.rogerscorp.com/documents/730/acm/ULTRALAM 3000 LCP laminate data sheet ULTRALAM 3850.aspx ory and Microwave Wiley IEEE Press 2005. [21]V .Veselag o, Soviet Physics Uspekhi, vol.10,no.4, pp. 509 515, Jan.,1968 [22] D. R. Smith, W. J. Padilla, D. C. Vier, S. C. Nemat Nasser, and S. Schultz. "Composite medium w ith simultaneously negative permeability and permittivity," Phys. Rev. Lett., vol. 84, no. 18, pp. 4184 4187, May 2000. [23] Kai Chang, Microwave ring circuits and related structures Wiley and sons, New York 2004. onductors an IEEE trans. Microw. Theory. Tech. Vol. 47, 1999, pp. 2075 2084 handed metamaterials: detailed numerical studies of the transmission properties J. Opt. A:Pure and Appl. Opt. Vo l.7 2005 pp.S12 S22 and right handed transmission peaks near the magnetic resonance frequency in composite metamaterials 2625 Circuit Models for Split Ring Resonators and Complementary Split Ring Resonators Coupl I IEEE trans. Microw. Theory. Tech. VOL. 53, NO. 4, APRIL 2005 [28] Justyna K. Gansel, et s Broadband Circular Science 18 September 2009, Vol. 325 no. 5947 pp. 1513 1515 [29] L. Solymar, E.Shamonina, Waves in Metamaterials Oxford University Press, 2009 [30] Pavel Kolinko and David R. Smith study of electromagnetic waves interacting with negative index materials 648

PAGE 162

162 [31] Mustafa K. Taher Al Filter Using SRR and CSSR: Design, Simu Antennas and Propagation (EuCAP), 2010 Proceedings of the Fourth European Conference on, 2010, pp.1 5 handed media simulation and transmission of EM waves in subwavelength split ring resonator loaded metallic wa veguides 2002, pp.183901 1 4. tive permeability IEEE Trans. Ant. Prop. Vol. 53, 2005, pp110 119. ic funnels for subdiffraction light compression and propagation 1 5 [35] David M. Pozar, Microwave Engineering Wiley, 1998. [36] Mingxu Liu; Craninckx, J.; Iyer, N.M.; Kuijk, M.; Barel, A.R.F.; "A 6.5 kV ESD pro tected 3 5 GHz ultra wideband BiCMOS low noise amplifier using interstage gain roll off compensation," Microwave Theory and Techniques, IEEE Transactions on vol.54, no.4, pp. 1698 1706, June 2006 [37]Reiha, M.T.; Long, J.R.; Pekarik, J.J.; "A 1.2 V r eactive feedback 3.1 10.6 GHz ultrawideband low noise amplifier in 0.13 /spl mu/m CMOS," Radio Frequency Integrated Circuits (RFIC) Symposium, 2006 IEEE pp., 11 13 June 2006 [38] C.A. Balanis, Advanc ed Engineering electromagnetics John Wiley & Sons 1989. circuit models for split ring resonators and complementary split Theory and Techniques, Vol.53, Issue 4, pp.1451 1461, Apr.2005. [40] W.J.R. Hoefer,"Equivalent Series Inductivity of a Narrow Transverse Slit in Microstrip," IEEE Transactions on Microwave Theory and Techniques, vol. 25, no. 10, pp. 822 824, Oct 1977 ent of Early Colo World Journal of Surgery, vol.21, no.7, 1997, pp. 694 701. endoscopy, vol.38, 1992, pp. 568 570. IF 1a and HIF 2a) Expression in Early Esophageal Cancer and Response to Photody Cancer Research, vol.61, Mar.2001, pp.1830 1832. vol 405, no. 6785, pp. 417 418, 2000.

PAGE 163

163 Know American Journal of Gastroenterology, vol.101, pp.2218 2222, 2006. [45] Chen, Zhining, Antennas for Portable Devices John Wiley and Sons, 2007. [46] Hoefer, W.J.R.; "Equivalent Series Inductivity of a Narrow Transverse Slit in Microstrip," Microwave Theory and Techniques, IEEE Transactions on vol.25, no.10, pp. 822 824, Oct 1977. [47] of Journal of the American Dental Association, vol. 142, no.2, pp. 170 172, Feb.2011. [48] http://www.omnivision.com/ [4 9 ] Gavish A. et al., temporomandibular Journal of Oral Rehabilitation, vol.27, pp. 22 32, 2000. [ 50 nce Allow 243, 2008. [5 1 and body position on tongue pressure with res Am J. Orthod Dentofacial Orthop., vol.115, no.3, pp.239 246, Mar.1999. [5 2 Journal of Oral Rehabilitation, vol.35, pp. 567 571, 2008. [5 3 opment of wireless bruxism monitoring device based on pressu re Sensors and Actuators A, vol.163, pp. 486 492, 2010. [5 4 vivo monitoring of bruxism with an intelligent tooth spl int IFMBE Proceedings, vol.25, no.11, pp.108 111, Sep. 2009. [5 5 R.; "Polymeric piezoelectric sensors and remote communication for detection of bruxism," Industrial Technology (ICIT), 2010 IEEE International Conference on pp.268 273, 14 17 March 2010 [5 6 sed on Polymeric Piezoelectric Sensors and Remote [5 7 ] Andres Diaz Lantada, Handbook of Active Materials for Medical Devices: Advances and Applications Pan Stanford Publishing, Singapore, 2012 pp.199 207.

PAGE 164

164 [5 8 splin The Journal of Prosthetic Dentistry, vol.86, issue 2, pp. 195 202, Aug.2001. [5 9 powered by an active interrogator intended for intra splint use for the detection and recording of bruxism," Pervasive Computing Technologies for Healthcare, 2009. PervasiveHealth 2009. 3rd International Conference on pp.1 4, 1 3 April 2009. [ 60 ] Jung Ho Kim; McAuliffe, P.; O'Connel, B.; Diamond, D.; King Tong Lau; "Development of Bite Guard for Wireless Monitoring of Bruxism Using Pressure Sensitive Polymer," Body Sensor Networks (BSN), 2010 International Conference on pp.109 116, 7 9 June 2010. [6 1 ] Zhenhai Chen; Luo, R.C.; "Design and implementation of capacitive proximity sensor using microelectromechanical systems technology," Industrial Electronics, IEEE Transactions on vol.45, no.6, pp.886 894, Dec 1998. [62] Juan Domingo Baena, Jordi Bonach e, Ferran Martn, Ricardo Marqus Sillero,Francisco Falcone,Txema Lopetegi, Miguel A. G. Laso, Joan Garca Garca, Ignacio Gil,Maria Flores Portillo, and Mario Sorolla "Equivalent circuit models for split ring resonators and complementary split ring reso nators coupled to planar transmission lines," Microwave Theory and Techniques, IEEE Transactions on vol.53, no.4, pp. 1451 1461, April 2005 [6 3 Sensors and Actuators A: Physical, vol. 75, issue. 3, pp. 242 251, June 1999.

PAGE 165

165 BIOGRAPHICAL SKETCH Xiaoyu Cheng was born in 1983 in Hebei, China. He received his Bachelor of Engine ering degree with the major of t elecommunications from the Hebei University of Science and Technology, China in 2006 and his Master of Science degree from University at Buffalo, the State University of New York in 2008. He is now working toward his Ph.D. degree with the Multidisciplinary nano and Microsystems (MnM) Laboratory and the Interdisciplinary Microsystems Group (IMG) at the University of Florida. His research interests include RF circuits and wireless medical devices, metamaterials and its application to medical instruments. He has more than 30 peer reviewed publications and he was a recipient of the IEEE Antenna and Propagation Society Doctoral Research Award 2012 and was selected as an Outstanding International Student of the University of Florida in 2011 and 2012. He was also a recipient of Bird Technologies Group Fellowship Award in 2009 and 2010.