Integration of UWB and Wireless Pressure Mapping in ... - IEEE Xplore

2550

IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 57, NO. 10, OCTOBER 2009

Integration of UWB and Wireless Pressure Mapping in Surgical Navigation Mohamed R. Mahfouz, Senior Member, IEEE, Michael J. Kuhn, Student Member, IEEE, Gary To, Student Member, IEEE, and Aly E. Fathy, Fellow, IEEE

Abstract—Wireless technologies are becoming more prevalent in hospital environments. An ultra-wideband (UWB) indoor tracking system is outlined, which has dynamic 3-D real-time root-mean-square error in the range of 5.24–6.37 mm using line-of-sight signals in different experiments. This high 3-D accuracy opens up many new applications to UWB indoor wireless positioning, which includes its use for tracking smart surgical tools and its ability to register relevant objects in the scene such as a spacer block for real-time pressure mapping of the femoral condyles. Experimental results quantifying the operating room (OR) environment for UWB transmission fit to the IEEE 802.15.4a channel model are included. A simulation of our UWB positioning system with interference from an IEEE 802.11a source shows the need for a front-end bandpass filter at our UWB receiver. Both microcantilever and microelectromechanical systems-based wireless pressure sensors are presented including quantitative performance metrics (e.g., hysteresis, sensitivity). The data for these sensors is transmitted over the 315- and 433.92-MHz telemetry bands. These bands are examined for their performance in the OR. Index Terms—Indoor positioning, microelectromechanical systems (MEMS), surgical navigation, telemetry, ultra-wideband (UWB).

I. INTRODUCTION

W

IRELESS technology has grown in the medical industry both in its prevalence and importance. Varshney describes current and future uses of various wireless technologies for medical applications including intelligent mobile emergency response systems, ad hoc networks combined with wireless local area networks (WLANs) for patient monitoring both inside and outside the hospital, the use of smart phones, and other portable devices equipped with 802.11 WLAN technology for telemedicine applications (i.e., everyday hospital use such as downloading daily schedules, patient information, etc.), and indoor tracking of assets, personnel, and patients through differential global positioning systems (GPSs) or RF identification (RFID) technology [1]. Pattichis et al. gives a comprehensive overview of the wireless technologies used in telemedicine, as well as an overview of systems deployed by various research Manuscript received February 20, 2009; revised June 25, 2009. First published September 15, 2009; current version published October 14, 2009. M. R. Mahfouz, M. J. Kuhn, and G. To are with the Mechanical, Aerospace, and Biomedical Engineering Department, University of Tennessee, Knoxville, TN, 37996 USA (e-mail: [email protected]; [email protected]; [email protected]). A. E. Fathy is with the Electrical Engineering and Computer Science Department, University of Tennessee, Knoxville, TN 37996 USA (e-mail: fathy@eecs. utk.edu). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TMTT.2009.2029721

groups to test the effect of telemedicine in various clinical settings for transmission of electrocardiograph, blood pressure, temperature, medical images, and other relevant signals for applications such as remote monitoring, teleradiology, and emergency care [2]. The use of implantable in vivo sensors, which utilize a telemetry system to transmit data from the body, have become more common with advances in science and technology. Evans et al. and To et al. have done extensive simulation, design, and experimental testing of microelectromechanical systems (MEMS) capacitive pressure sensors and microcantilever sensing technology for recording a pressure map of the knee stresses on the spacer block used in a total knee arthroplasty (TKA) incorporating RF telemetry in the 315- and 433-MHz bands [3], [4]. Pritchard et al. extended this sensor technology for in vivo use by embedding the sensors in a parylene coating for placement inside the polyethylene insert of a TKA and optimizing sensor performance for in vivo loading, which allows pressure mapping of the knee stresses on the femoral condyles [5]. Computer-aided surgery (CAS) has played an increasingly important role in the positive outcome of orthopedic surgeries since gaining widespread prevalence and acceptance in the 1990s [6]. The main type of CAS system used in orthopedic surgeries is a passive system, and integral to these systems are a tracking capability to give real-time feedback to the computer and surgeon on the 3-D position and orientation of surgical tools and patient anatomy [7]. Optical and electromagnetic tracking systems are the de facto standard in orthopedic CAS. Both of these systems provide 6 degrees-of-freedom (DOF) real-time feedback with sub-millimeter translational accuracy and sub-degree rotational accuracy (e.g. Polaris Spectra, Northern Digital Inc., Waterloo, ON, Canada [8] and 3-D Guidance medSAFE, Ascension Technology Corporation, Milton, VT [9]). Both optical and electromagnetic tracking systems have certain limitations when used for 3-D indoor tracking. Optical systems require a line-of-sight (LOS) signal between the probe and optical receiver, as shown in Fig. 1, while electromagnetic systems have a limited view volume and are susceptible to metallic interference (although new technologies, such as those employed in the 3-D Guidance medSAFE system, provide much better performance in terms of 3-D accuracy in the presence of metallic interference compared to previous generation systems). Both of these systems have been tested under various static and dynamic conditions typically observed during orthopedic CAS with results deemed satisfactory for CAS, assuming none of these limitations are violated [10]. Wireless tracking systems have never been seriously considered in orthopedic CAS because the accuracy required in

0018-9480/$26.00 © 2009 IEEE

MAHFOUZ et al.: INTEGRATION OF UWB AND WIRELESS PRESSURE MAPPING IN SURGICAL NAVIGATION

2551

TABLE I SUMMARY OF LICENSED MEDICAL WIRELESS FREQUENCY BANDS

Fig. 1. Optical tracking system used in orthopedic surgical navigation.

these applications is on the order of 1–2-mm 3-D translational error and 1 –2 3-D rotational error. For example, Clarke et al. tested five commercial systems employing Wi-Fi, received signal strength, RFID, ultrasound, and ultra-wideband (UWB) technology [11]. The systems were tested in multiple locations inside an operating room (OR) and compared for their performance in 3-D indoor localization. UWB performed orders of magnitude better than the other technologies with average 3-D accuracy near 10 cm (using the Ubisense, Cambridge, U.K., real-time location system [12]). The real-time location system from Ubisense [12] and the Sapphire DART system from Multispectral Solutions Inc., Germantown, MD [13] are commercially available and have similar operating ranges ( 50 m), frequency bands of operation (5.8–7.2 GHz), and 3-D real-time accuracy (10–15 cm). Frequency modulated continuous wave (FMCW) localization systems offer a complementary positioning technology and have been well documented both in research and commercially with dynamic accuracy from 1 to 10 cm and operating ranges as high as 400 m [14]–[17]. Meier et al. reported static 1-D accuracy of 0.1 mm for a 24-GHz UWB system with the transmitter and receiver wired together incorporating a Kalman filter while dynamic 3-D accuracy reduced to 2 mm [18]. In this paper, a novel UWB indoor positioning system is introduced [19]–[25], and its use in orthopedic CAS is discussed including its incorporation into smart surgical tools (e.g., a wireless probe for real-time bone morphing) and its use in a robust telemetry capacitive sensor system used in the ligament balancing performed during a typical TKA. Section II introduces the wireless technologies used both for positioning (UWB) and telemetry. Experimental results which quantify the OR environment for UWB transmission are included in Section II-A. The 315- and 433-MHz telemetry bands are examined in Section II-B including operating range and a simulation to examine the losses incurred when transmitting through the soft tissues of the knee. Section III introduces the complete surgical navigation system. It is broken into pre-operative, intra-operative, and post-operative phases. Sections III-C and D discuss the microcantilever and MEMS-based capacitive pressure sensors. Section IV presents the UWB positioning

system including dynamic and static 3-D real-time results. Section IV-B examines the performance of the system in the presence of an IEEE 802.11a WLAN interferer. Finally, Section V presents a conclusion. II. WIRELESS TECHNOLOGY OVERVIEW Both the wireless telemetry system used in the spacer block and the UWB positioning system have stringent requirements on the frequency bands within which they can operate. Table I highlights the different bands both in the U.S. and Europe, which can be used for indoor medical applications for both narrowband and UWB applications. UWB has available frequency bands from 3.1 to 10.6 and 22 to 29 GHz in the U.S. Only portions of that 3.1–10.6-GHz band are currently available in Europe. A number of telemetry bands exist in the U.S., and both the U.S. and Europe have instrumentation, scientific, and medical (ISM) bands available, mainly in the 300-MHz–3-GHz range. The UWB positioning system operates from 5.4 to 10.6 GHz in the upper region of the 3.1–10.6-GHz band, while the spacer block telemetry system operates at 433.92 MHz in the 433. 05–434.79 European ISM band with a proposed parallel design operating at 315 MHz for the U.S. ISM band. A. UWB in the OR The main concern with using wireless tracking technology in the OR is the high level of scatterers and corresponding multipath interference experienced when transmitting wireless signals. While the experiment from Clarke et al. [11] provides quantitative data on how wireless real-time positioning systems perform in the OR, it is also useful to look into UWB channels and their effect on UWB signals and 3-D tracking. There are two typical approaches used when modeling UWB channels: the first is statistical models used to model generic environments (e.g., industrial, residential, commercial, etc.), which

2552


incorporate LOS or non-line-of-sight (NLOS) measurements taken in the time and frequency domains, which are then used in setting the parameters of these statistical models. The second method uses ray tracing techniques to model specific geometrical layouts (e.g., buildings, cities) and can provide a more accurate depiction of which obstacles and structures will have the greatest effect on wireless propagation [26]. The drawback with ray tracing is the static nature of the results (i.e., results are only valid for a certain scenario of objects placed in the scene). Even if the base stations in the OR are static, other objects such as people, patients, operating table, medical equipment, etc. will not be. Therefore, in developing our UWB positioning system, we have focused on the use of the IEEE 802.15.4a channel model, which uses a large number of parameters and random variables in modeling a wide variety of environments (e.g., indoor industrial LOS/NLOS, indoor commercial LOS/NLOS, indoor residential LOS/NLOS, outdoor, body-area networks) [27], [28]. The impulse response of the UWB channel in the time domain is shown in (1), while the pathloss model used in the corresponding UWB channel is shown in (2). These models have been used both in simulating the behavior of our UWB positioning system [22], [23] and in designing and implementing

TABLE II SUMMARY OF PARAMETERS FIT TO IEEE 802.15.4A CHANNEL MODEL WITH EXPERIMENTAL UWB DATA TAKEN IN THE OR

(1) (2) our receiver-side pulse detection algorithm for highly accurate 1-D distance measurements between the transmitter and receiver, integral to the high performance of our system [19],[26]. Although the IEEE 802.15.4a UWB channel model has been fit to many environments, the OR is one place where measurement data is lacking and modeling information is still nonconclusive. The most comprehensive analysis was done by Hentila et al. who used time- and frequency-domain measurement techniques in analyzing UWB channels in different hospital environments including the OR, X-ray examination room, and intensive care unit [29]. The data best fit to a Nakagami-m distribution using a modified Saleh-Valenzuela model, similar to (1). We performed extensive time- and frequency-domain measurements in the OR both during surgery (live) and not during surgery (nonlive) for Tx–Rx distances of 0.5–4 m. Table II shows a truncated list of parameters for the LOS OR environment fit to the IEEE 802.15.4a channel model, which were obtained with time- and frequency-domain experimental data. Fig. 2 shows the pathloss for the OR environment obtained by fitting experimental data to (2) and compared to residential LOS, commercial LOS, and industrial LOS. The pathloss in the OR is most similar to residential LOS, although this can change depending on which instruments are placed near the transmitter and receiver. Fig. 3 shows path loss obtained for a Tx–Rx distance of 0.49 m where the transmitting (monopole) and receiving (Vivaldi) antenna effects have been removed. Small scale fading effects can be seen as well as frequency-dependent pathloss, which is captured in the parameter in Table II. Fig. 4 shows an example time-domain received signal for a Tx–Rx

Fig. 2. Comparison of pathloss for IEEE 802.15.4a LOS channels. The pathloss for the OR environment is most similar to residential LOS.

distance of 1.49 m using the monopole antenna for transmitting and single-element Vivaldi antenna for receiving. A decaying exponential is overlayed on the received signal to highlight the in Table II. The pathloss of intra-cluster decay, defined by the LOS OR channel is most like a residential LOS environment whereas the power delay profile (PDP) is closer to an industrial LOS environment where dense clusters of multipath quickly decay. Fig. 5 shows electromagnetic interference (EMI) measured in a standard dual OR during four orthopedic surgeries [30]. EMI is prevalent below 3 GHz with technologies such as CDMA-2000, U.S. Cellular Band, Bluetooth, and WLAN all residing in this frequency range. Our UWB positioning system operates over a frequency band of 5.4–10.6 GHz, as shown in Fig. 6, effectively bypassing all potential EMI, except for the U.S. ISM band at 5150–5875 MHz, which is where IEEE 802.11a WLAN systems currently operate (802.11n systems also operate in this band). Since interference from this band is a concern for our


Fig. 3. Pathloss obtained with the Tx and Rx placed 0.49 m apart where effects from the transmitting (monopole) and receiving (Vivaldi) antennas have been removed. The frequency dependence, in Table II, can clearly be seen, as well as small scale fading effects.

2553

Fig. 5. EMI measured in the OR during four orthopedic surgeries from 200 MHz to 26.5 GHz.

0

Fig. 4. Example received signal in the time domain for a Tx–Rx distance of 1.49 m highlighting the distortion (seen as expansion) in the LOS pulse due to a dense cluster of multipath rays. The overlayed exponential is fit using , as outlined in Table II, to show the intra-cluster decay of the LOS cluster.

system operating in the OR, a comprehensive simulation of potential WLAN interference is outlined in Section IV. As mentioned in Section IV, the use of a bandpass filter at the front-end of our UWB receiver for each base station effectively mitigates WLAN interference, even when the WLAN transmitter is only 1–2 m away from the base station and transmitting a LOS signal. B. 315- and 433-MHz Wireless Telemetry in the OR Attenuation of the wireless signal due to soft tissue is one of the major obstacles in designing in vivo wireless systems, such as the wireless telemetry pressure sensor described in Section III-C, which operates at either 315 or 433 MHz. The lossy nature of soft tissue to signals in the RF and microwave frequency range has been extensively researched. The electrical

Fig. 6. Power spectral density of simulated UWB signal showing 10-dB bandwidth from 5.4 to 10.6 GHz, which bypasses interfering signals observed in Fig. 5. This spectrum contains the IEEE 802.11a WLAN band (5.15–5.875 GHz). Section IV-B discusses in detail how to mitigate this interference in our UWB positioning system.

properties of an arbitrary medium (e.g., muscle, fat) can be and completely characterized by the relative permittivity conductivity (3) where is the complex permittivity, is the imaginary portion of the complex permittivity signifying loss, and is the radial frequency. Gabriel et al. [31] performed open-ended coaxial probe measurements on a broad spectrum of human tissues at 37 C and developed parametric models of the complex permittivity (4) where each term in the summation represents a separate Cole–Cole dispersion region and the last term accounts for

2554


TABLE III SIMULATED LOSSES FOR TRANSMITTED ASK SIGNALS AT VARIOUS FREQUENCIES

Fig. 9. Power spectrum measured intra-operatively during a knee replacement surgery [30] from 200 MHz to 2.5 GHz.

Fig. 7. Electrical properties of human muscle, fat, and skin at 37 C. (a) Relative permittivity. (b) Conductivity.

Fig. 8. Overview of simulation setup where the soft tissues of the knee are simulated using a 10-mm layer of muscle, 5-mm layer of fat, and 1-mm layer of skin followed by air.

ionic conductivity. Fig. 7 shows the relative permittivity and conductivity for human fat, muscle, and skin at 37 C over a large frequency range. It is clear that the conductivity, and therefore, the loss, begins to sharply rise for all three tissues starting around 800 MHz. Fig. 8 outlines a simulation conducted using the tissue properties in Fig. 7. The electric field and tissue interfaces were simulated following the procedure outlined in [32] where a recursive process is used to solve for the equivalent impedances of each tissue and reflection coefficients for each tissue interface (e.g., muscle–fat, fat–skin, and skin–air). The signal first travels

through 10 mm of muscle and then through the muscle–fat interface, then through 5 mm of fat and through the fat–skin interface. Part of the signal is reflected at the fat–skin interface, which then travels back to the fat–muscle interface. Solving for the equivalent impedances and reflection coefficients for each tissue and interface in the same order as the emitted signal takes into account this phenomenon where electric field waves can effectively bounce around between two interfaces [32]. For the simulation, a modulated amplitude shift-keying (ASK) signal is transmitted at 10-dBm power at 315, 433, and 915 MHz and 2.4 GHz. As shown in Table III, all signals will experience some loss due to the tissues. The losses incurred at 315 and 433 MHz are 12.5–22.5 dB lower than the losses incurred at 915 MHz and 2.4 GHz. Operating at 315 and 433 MHz successfully meets the operating range of our system while operating at 915 MHz and above results in losses through the medium too high to obtain an operating range of even 5 m. It is necessary to examine the susceptibility of this wireless system to interference from other sources. Fig. 9 shows the measured power spectrum in the OR from 200 MHz to 2.5 GHz [30]. Signals from CDMA-2000 are noticeable at 872 and 928 MHz, while a U.S. cellular band signal can be seen at 1.95 GHz and a WLAN signal at around 2.4 GHz. No interference was observed at 315 or 433 MHz in this experiment. It is worth noting that RFID systems are commercially available for hospital use which operate in the 433-MHz band [33]. Also, electrocardiogram (ECG) telemetry systems have been designed, which operate in the 433-MHz band using frequency


2555

Fig. 10. Overview of orthopedic surgical navigation system including pre-operative, intra-operative, and post-operative phases.

shift keying (FSK) at ranges up to 40 m [34]. As mentioned in [35], for the ASK transmitter and receiver employed in our system, 30–40 dB more pathloss is noticed indoors versus outdoors. This significantly reduces the operating range of systems using the 315- and 433-MHz bands inside a hospital. For example, given our receiver sensitivity of 115.6 dBm, our transmitter power of 10 dBm after the chip antenna (discussed in Section III), the tissue losses outlined in Table III, and the indoor pathloss specified by the manufacturer [35], the maximum operating range of our telemetry system is 20 m in the 315-MHz band and 15 m in the 433-MHz band, both satisfying our operating range requirement of 5 m. Each transmitted packet includes a 16-bit header, which employs even parity checking and a unique bit sequence for identification. This is followed by 270 bits for the 30 channels of microcantilever data, which employs odd parity checking to ensure proper transmission. The use of parity checking and a unique header in data transmission helps mitigate effects from interference seen at the 315- and 433-MHz bands. III. SURGICAL NAVIGATION OVERVIEW The main purpose of surgical navigation systems is to provide real-time feedback and quantitative assessment to the surgeon regarding the quality of the surgical process. The overall surgical process can be divided into three distinct steps, which are: 1) pre-operative; 2) intra-operative; and 3) post-operative. Fig. 10 outlines how novel technologies can be used in each of these steps to help provide more information with better accuracy for surgeons, implant designers, and patients. In the pre-operative step, technologies such as automatic 3-D segmentation of computed tomography (CT) datasets and biplanar X-ray reconstruction are used in creating patient-specific 3-D models of the patient’s bones, which are then tracked on a computer in the intra-operative phase [36], [37]. Once an accurate model of the patient’s anatomy is created, advanced analysis of the joint anatomy can be used to provide feedback to the surgeon on implant sizing and other relevant specifications, as shown in Fig. 11 [38]. In the intra-operative phase, 3-D real-time tracking of the patient’s bones, as well as surgical tools, will be done through registering a bone tracking system with our UWB positioning system and using the UWB positioning system for tool

Fig. 11. User interface for integrated software application, which performs virtual resection of a patient-specific bone using a statistical bone atlas and advanced morphometric measurements. It is also used for biplanar X-ray reconstruction and bone morphing.

tracking, tracking of the capacitive sensor ligament balancing system into the surgical scene, and finally incorporation of the UWB technology into smart surgical tools. The ligament balancing system uses microcantilever pressure sensors. The smart provisional, used to determine patient implant size intra-operatively, uses MEMS-based pressure sensors. The final phase is post-operative analysis. A novel 3-D to 2-D registration algorithm is already in use for extracting 6 DOF (3-D translation and 3-D rotation) information from fluoroscopic images [39]. This will be combined with a gait analysis laboratory. Current gait analysis is done by either tracking with one or more video cameras or by using passive or active optical markers, which reflect or emit signals to a multiple (e.g., 8) of infrared cameras. Another portion of the post-operative step is the smart implant, discussed in Section III-D, where MEMS capacitive sensors are embedded in a polyethylene component in a knee im-

2556


Fig. 12. Illustration of a smart surgical tool. The spacer block, with an embedded pressure sensing-telemetry system, is tracked in 3-D by an UWB localization probe.

Fig. 13. Illustration of the MEMS spacer block integrated with the UWB positioning system for 3-D real-time tracking. UWB probes are rigidly fixed to the femur and tibia to track the bones in real time. A UWB probe is also rigidly attached to the spacer block to track it in real time.

plant to give internal pressure readings post-operatively. Fig. 10 provides an overview of the complete navigation system. More details are provided on the intra-operative step including smart surgical tools in Section III-A, the post-operative step including gait analysis in Section III-B, the microcantilever ligament balancing system in Section III-C, and the MEMS capacitive pressure sensors in Section III-D. A. Intra-Operative Phase As shown in Fig. 10, the intra-operative phase consists of real-time tracking of smart tools or instruments, bones, and the capacitive ligament balancing system. The UWB tracking system is covered in detail in Section IV. The microcantilever ligament balancing system is covered in Section III-C. The MEMS capacitive pressure sensors, which are used in smart provisionals to determine the implant size for the patient intra-operatively, are discussed in Section III-D. The smart provisional provides intra-operative feedback to optimize soft tissue interaction with orthopedic joint replacements. The smart provisional system utilizes the 315- and 433.92-MHz frequency bands for data transmission. The incorporation of the intra-operative pressure sensing system inside the spacer block combined with an UWB localization probe that tracks its location in real time, as illustrated in Fig. 12, is an example of the future technology in surgical navigation: smart surgical tools. Fig. 13 further illustrates the integration of the UWB tracking system and microcantilever pressure sensing spacer block system. A UWB probe containing a 3–6-element UWB monopole array is used to obtain the 3-D

Fig. 14. (a) Capacitive sensor placed on implant used in TKA. (b) Fluoroscopy image with 3-D implant models overlayed. (c) Contact areas and pressures can be compared between fluoroscopic analysis and real-time pressure mapping of the condyles via sensor arrays.

position and 3-D orientation of a rigid body. The femur, tibia, and spacer block have UWB probes rigidly attached to them, as shown in Fig. 13. As discussed in Section IV, the UWB positioning system has static 3-D accuracy of 4.67 mm and dynamic 3-D accuracy of 5.24–6.37 mm. This can cause 4–6 rotation error for mm of translation error and 1 –3 of each rigid body. As discussed in Section IV, this is still too high for surgical navigation applications where 1–2 mm of rotation is needed 3-D translation accuracy and 1 of to ensure proper implant alignment to the patient anatomy. Conversely, this accuracy is acceptable for ligament balancing applications since the system can still use the pressure mapping of the medial and lateral condyles to monitor the compartmental pressures. The UWB positioning system provides quantitative data on the varus/valgus condition of the knee. It also provides an accurate measure of the quality of the distal and posterior cuts on the femur to the proximal cut on the tibia. This quickly alerts the surgeon to any problems in the surgical resection intra-operatively. B. Post-Operative Phase The smart implant provides information about the post-operative loading within the joint. Fig. 14(a) demonstrates the flexibility of the MEMS sensors as they conform to the shape of the polyethylene implant. Fig. 14(b) shows a fluoroscopy image of an implanted knee with 3-D models overlayed for the femur and tibia implants and also the polyethylene insert. Fig. 14(c) shows the corresponding pressure map of the smart implant components. The smart implant system utilizes the 315- and 433.92-MHz frequency bands for data transmission. Section III-D provides more details on the MEMS capacitive pressure sensors. Finally, Fig. 15(a) shows a post-operative analysis system where a hybrid system combines tracking data, from either an optical or UWB-based tracking system, with image data [26]. The data is fused to provide highly accurate gait analysis. Fig. 15(b) shows a conventional gait positioning system incorporating optical markers [26]. UWB tracking has the potential to integrate with sensor-based systems and even medical imaging modalities to create advanced intra-operative and post-operative guidance systems with the potential to revolutionize the technology used in surgical navigation.


2557

Fig. 15. Post-operative analysis. (a) Hybrid system combines tracking systems with imaging modalities for tracked visualization and information fusion. (b) Positioning system used for post-operative gait laboratory analysis.

Fig. 17. Block diagram of microcantilever pressure telemetry system [41].

Fig. 16. (a) Simple resection gap assessment tool. (b) Wired strain mapping tool (on right side).

Fig. 18. (a) Microscopic view of single microcantilever. (b) Fabricated ASIC for signal processing of microcantilever capacitive sensor [4].

C. Microcantilever Ligament Balancing System Incorporating sensors on standard surgical instrumentation during orthopedic surgery is not a new goal for improving surgical outcomes, but it remains elusive in implementation due to difficulty sensing enough information in a relevant timely manner to display to the surgeon. One of the most important factors to ensure the survival of knee joint implants is the intra-operative soft tissue balancing during TKA. The force must be enough to keep the knee stable, but not too much to induce large pre-stress at the joint surface. Both tibial insert height and ligament tightness contribute to this passive force. Surgeons rely on their experience or simple qualitative surgical instruments to assess the amount of femoral and tibial resections , as shown in Fig. 16(a) and (b), where an assessment gap measuring tool and strain mapping tool are used in the ligament balancing. A quantitative approach is to map the stress between the femur and tibia intra-operatively. The challenges with developing a pressure mapping instrument include the size of the sensors, the electronics, and the wires connecting to the data acquisition and analysis system. Previous research shows that strain mapping devices require long and cumbersome wires, which are not desirable in the OR [40]. A microcantilever-based pressure sensing system was previously developed [4]. Fig. 17 shows a block diagram of the pressure sensing system, while Fig. 18(a) shows a microcantilever and Fig. 18(b) shows a fabricated application-specific integrated circuit (ASIC) used for signal processing of the microcantilever sensor array [4]. Table IV shows the properties of a microcantilever embedded in 2 mm of EP30MED epoxy. Pressures from 0 to 300 kPa can be measured in the linear range. Pressures as high as 3 MPa can be measured. The linearity of the microcantilever

TABLE IV PROPERTIES OF MICROCANTILEVER ENCAPSULATED IN 2-mm OF EPOXY (EP30MED)

is 0.625 mV/kPa. The sensitivity or resolution of the microcantilever is 0.355 mV/kPa. The repeatability is 0.644 mV/kPa over the pressure range of 0–300 kPa. Pressures above 300 kPa are common in the knee joint. Therefore, it is useful in post-operative applications, such as the smart implant, to be able to measure pressures above 300 kPa in the linear range. The MEMS capacitive pressure sensing system can operate at pressures above 2 MPa and is outlined in Section III-D. After applying dc power to the microcantilevers, a 5-min warm-up period is required for proper sensor operation. After encapsulation of the microcantilevers in epoxy, some hysteresis can be observed over an extended period of time. Recording this behavior for various pressures allows the effects of hysteresis to be effectively calibrated out of the final pressure readings. After sending the microcantilever signal through the multiplexer, amplifier, and ADC, the sensitivity of the system in measuring pressure differences is 1.79 kPa. After encapsulating the microcantilevers in epoxy, the resonant frequency of the sensor is reduced below 20 Hz and its effect on the output data is negligible. As mentioned in Section II-B, the RF signal attenuation of the soft tissue limits the choice of the wireless frequency band. For

2558


Fig. 20. Fabricated upper layer of microcantilever pressure sensing system including chip antenna for telemetry of data for use inside a spacer block for highly accurate pressure sensing across both the medial and lateral femoral condyles.

Fig. 19. Fabricated: (a) transmitter and (b) receiver for the microcantilever pressure sensor system.

in vivo and intra-operative wireless environments, the system is operating within the 315- and 433.92-MHz bands. Hence, the pressure sensing system comes equipped with a chip antenna operating at 433.92 MHz, which allows it to transmit data even when inserted into the knee joint of the patient, as discussed in Section II-B. The transmitting and receiving boards for the microcantilever pressure sensing system are shown in Fig. 19. A chip antenna from Tricome is used for telemetry at 433 MHz (European ISM band). The transmitter and receiver used are the MAX1472 and MAX1473 chips from Maxim-IC Inc., which use ASK modulation. The data rate is approximately 100 kb/s. However, the transmitter power drops to 10 dBm after switching to the chip antenna, as compared to 10 dBm, as stated in the specification for a normal monopole antenna. As mentioned in Section II-B, the sensitivity of the wireless system operating at 433.92 MHz is 115.6 dBm with approximately 0.2% BER. This provides a 15-m operating range when transmitting through tissue, surpassing the 5-m requirement needed by our surgical navigation system. Fig. 20 shows the fabricated upper layer of the microcantilever pressure sensing system. D. MEMS Capacitive Pressure Sensors The MEMS technology, which can be used to measure pressures in the knee joint similar to the microcantilever technology outlined in Section III-C, is shown in Fig. 21. It is mentioned in

Fig. 21. Single MEMS tri-axial pressure sensing element on SiO .

[3] and has been extended in its use as a smart provisional and smart implant, as outlined in [4] and [5]. These concepts are enabled with flexible capacitive biocompatible MEMS technology to give real-time feedback on the intercompartmental pressures within the knee joint during motion. A MEMS tri-directional pressure sensing element is shown in Fig. 21, which enables the monitoring of axial and shear forces experienced by the material. All strain tensors can be discerned, as given by (5)

The sensor design is solid state; strain within the dielectric accounts for a change in capacitance, which is monitored in real time. This makes the sensor robust for repetitive loading and ideal for intra-operative and implant use. The sensing element is comprised of six sensors, with one sensor optimized for response to axial strain, two interdigitated sensors for in-plane strain, and a three-sensor rosette of differential sensors for response to shear strains. Sensor optimization was completed through finite-element simulation with Coventor (Cary, NC) [42]. Elements having nominal capacitances of 1, 5, and 10 pF were microfabricated and tested using a hydraulic MTS858 table top mechanical testing machine with 2.5-kN load cell. The test consisted of alternating between no load and a sequentially


2559

then sampled with a conventional ADC (105 MSPS, 10 bit) and sent to a field-programmable gate array (FPGA) where a novel leading-edge detection algorithm locates the pulse leading-edge with 1–2 mm of accuracy, even in dense multipath environments [26]. More information on the system architecture, including the leading-edge detection algorithm, use of and , the unique single channel an offset between noncoherent approach, and the use of a sub-sampling mixer can be found in [19], [24], and [26]. Section IV-A discusses real-time experimental results from the UWB positioning system including accuracy, hysteresis, and drift, and how these affect its performance in tracking smart instruments and bones intra-operatively. Section IV-B discusses the effects of WLAN interference in the 5.15–5.875-GHz band and how these effects can be mitigated by using a bandpass filter at the front-end of our UWB receiver architecture, shown in Fig. 23. Fig. 22. Load profile for capacitance array test for a 5-pF sensor within an array. Sensor response is linear with respect to axial load and shows minimal hysteresis.

increasing load with a return to 0 N. All loads were held for 2 min, and the loading and unloading rate was at 50 kPa/s. This test was used to measure static and dynamic loading conditions, repeatability, and drifting of the MEMS sensor. Fig. 22 shows the time series data from the loading test for a representative 5-pF capacitor. The 10-pF capacitor has a sensitivity of 2.86 10 pF/MPa, the 5-pF capacitor has a sensitivity of 1.54 10 pF/MPa, and the 1-pF capacitor has a sensitivity of 4.93 10 pF/MPa. With loading normal to the sensing element, as the nominal capacitance decreases, the sensitivity improves. For the shear sensors within the sensing element, a differential measurement is taken so nominal capacitance is not critical. For axial loading, hysteresis of less than 3% was observed in this mechanical testing, with negligible drift when properly shielded. Pressures of greater than 2 MPa can be measured while still operating in the linear range of the sensor. After amplification of the capacitive signals and conversion to digital with a 24-bit ADC, the final calculated sensitivity of the system is 30 Pa for normal loading. A modified test setup was used to submit the sensing element to primarily shear loading. A balanced design was used to exercise tensile force along the test rig to produce shear loads at the plane of the sensing element. The sensor exhibited a response to shear loads of 4.77 10 pF/MPa. Finally, in-plane loads were isolated in a test rig, showing linear response to increased in-plane loads. IV. UWB POSITIONING SYSTEM The UWB positioning system is outlined in Fig. 23. More information on the UWB positioning system can be found in [19]–[24], [26]. A Gaussian pulse is modulated by an 8-GHz carrier signal and transmitted via an omni-directional UWB antenna. The signal is received at the base stations by a single-element Vivaldi antenna, goes through two stages of low-noise amplification, is downconverted, and then one channel [not both in-phase (I) and quadrature (Q)] is sent to a sub-sampling mixer where it is time extended by sampling a pulse train of 500–10 000 pulses. The microsecond pulse is

A. Experimental Results The experimental setup used in testing our system is outlined in [19] and [24]. The use of an optical tracking system, which has higher accuracy than the UWB positioning system with 0.3-mm 3-D root-mean-square error (RMSE), is used and provides reference 3-D positioning data to compare the 3-D data of the UWB system. The optical tracking system used is the hybrid Polaris Spectra from Northern Digital Inc. [8]. Data from both the optical tracking system and UWB positioning system are synchronized and sent to a computer. In contrast to previous experiments, since this experiment was being run in real-time, a twofold calibration procedure is needed to calibrate the UWB system to the optical tracking system. First, the cable length offsets of the UWB system are calibrated by using the optical system as a reference and iterating through offset values used in the time differences of the time-difference-of-arrival (TDOA) calculation. Once the cable length offset calibration has occurred, a full 3-D calibration is done where a 3-D transformation matrix is obtained to calibrate the UWB system with the probe tip of the optical passive probe. This is done by moving the combined apparatus containing the optical and UWB systems around the 3-D view volume and collecting points across the entire view volume for use in the final 3-D calibration [26]. Once an accurate 3-D calibration has been carried out, synchronized real-time 3-D data is collected from both systems. Median inter-quartile-range (MIQR) filtering and averaging of time indices and TDOA values is needed to obtain millimeter accuracy. A summary of the experimental results for the UWB positioning system is shown in Table V. The system has been tested in both static and dynamic scenarios. The dynamic 3-D accuracy for experiments using free-form motion [24], robot-controlled motion [24], and an optical rail is 5.24–6.37 mm. This is still too high compared to 1–2 mm, which is needed in orthopedic surgical navigation systems to ensure correct 6 DOF tracking of instruments and bones. As shown in Table V, in static experiments, the accuracy is 4.67 mm for 20 distinct locations for a and noncoherent experiment where the transmitter uses [24]. In the coherent case where the transthe receiver mitter and receiver are run off the same clock, static 3-D accuracy of 0.73 mm is achieved. This is below the error threshold for orthopedic surgical navigation intra-operative tracking.

2560


Fig. 23. Block diagram of UWB positioning system outlining tag and base-station architectures, as well as post-processing on a computer.

UWB data after a prolonged time period (e.g. 30 min). This is caused by the instability of the noncoherent PRF clocks. As the expansion factor of the received pulse at the sampling mixer is increased, or more points are averaged on the computer, the effects of time scaling and drift become more noticeable. The initial calibration procedures and also the use of a reference tag in a fixed and known location is needed to help mitigate the effects from time scaling and drift. Using the robot experiments, hysteresis of the system was measured around 3%–5% in a 0.60 0.20 0.15 m view volume [26]. In the future, the full integration of the UWB system with smart instruments and bone morphing will solidify the use of UWB technology in high-accuracy indoor applications.

TABLE V SUMMARY OF 3-D REAL-TIME EXPERIMENTS

B. 5-GHz WLAN Interference

2

Fig. 24. Drift is noticeable when localizing a static point with 10 000 averaging over 1000 samples. The geometric position dilution of precision (PDOP) causes drift error in the x dimension to be amplified relative to y and z .

Fig. 24 shows data obtained using 10 000-point averaging for 1000 samples. Noticeable drift can be seen in the

The spectrum of our UWB positioning system is 5.4–10.6 GHz, as shown in Fig. 6. Since the ISM band extends from 5.15 to 5.875 GHz, and our UWB monopole and Vivaldi antennas operate over this range, the susceptibility of our system to interference from IEEE 802.11a/n systems needs to be addressed. We have simulated the performance of our system in the presence of a 5-GHz interferer utilizing the simulation framework in [23]. Agilent ADS was used to generate an IEEE 802.11a WLAN signal. The approach for adding channel effects to the IEEE 802.11a signal is similar to that followed by Bellorado et al. [43]. The pathloss model for the OR shown in Fig. 2 was used for modeling the pathloss of both the UWB and IEEE 802.11a signals. LOS between the IEEE 802.11a source and the UWB base stations was assumed. Multipath fading was modeled by a Rician fading channel with a K-factor of 6 dB for the WLAN system. Interference from the IEEE 802.11a source was examined at distances of 1, 2, 5, and 10 m from the UWB base stations. The distance from the UWB transmitter to each UWB base station was kept between 1–2-m LOS and the time-domain indoor channel for the OR described


2561

Fig. 25. Power spectral density showing the UWB spectrum where: (a) the UWB pulse from our positioning system can be seen, as well as interference from an IEEE 802.11a source placed 2 m from the UWB base station and (b) the IEEE 802.11a signal is filtered out by the UWB base station.

in Section II-A was used to model multipath interference for the UWB system. Fig. 25(a) shows the power spectral density for an input signal seen at our UWB receiver where both the UWB pulse and IEEE 802.11a signals are present. The WLAN signal is 2-m LOS from the UWB base station. The time-extended signal obtained by our UWB receiver for the received signal in Fig. 25(a) is shown in Fig. 26(b). No visible UWB pulse is detected because of the high-power WLAN interfering signal. Fig. 25(b) shows the received signal after passing through two stages of amplification and a three-stage stub-based bandpass filter where a 35-dB stopping amplitude is seen at 5.5 GHz and the passband begins at 6.8 GHz. Zhu et al. achieved comparable experimental performance with multistage bandpass filters using ground plane aperture techniques [44]. The added distortion of the UWB pulse from the bandpass filter does not affect the overall system accuracy since our leading-edge detection algorithm only looks for the initial leading edge rather than attempting to reconstruct the correct pulse shape. Fig. 26(c) shows the time-extended received signal for the power spectrum shown in Fig. 26(b). The UWB pulse is once again clearly seen, although, as shown in Table VI,

Fig. 26. Received time-domain signals after passing through the UWB receiver architecture in Fig. 23 including two stages of amplification, downconversion, low-pass filter (LPF), sampling mixer, and ADC conversion. (a) Time-extended UWB pulse with no 5-GHz interference. (b) Time-extended UWB pulse with 5-GHz interference where receiver saturation occurs. (c) Time-extended UWB pulse with 5-GHz interference and a bandpass filter to mitigate the effects of the IEEE 802.11a interferer.

with the addition of a bandpass filter to remove the WLAN interference, the signal-to-noise ratio (SNR) of the received pulse drops 6.2 dB for a WLAN interferer placed 2 m from the UWB base station. Similar results for the received SNR are seen for

2562


TABLE VI SNR OF RECEIVED TIME-EXTENDED UWB SIGNALS WITH AND WITHOUT IEEE 802.11A INTERFERENCE

ysis. The authors would like to thank E. Pritchard, University of Tennessee, for material related to the smart implant and MEMS technology. The authors would like to thank Dr. C. Zhang, Hittite Microwave Corporation, Chelmsford, MA, for his contributions in developing microwave hardware and antennas for the UWB positioning system. Finally, the authors would like to thank Dr. B. Merkl, Medtronic Navigation, Louisville, CO, for his contributions in developing digital processing techniques and TDOA algorithms for the UWB positioning system. REFERENCES

WLAN transmitting distances of 1, 5, and 10 m with the SNR increasing as the WLAN interferer moves farther away (9.74 dB with a 10-m distance). The use of automatic gain control (AGC), as discussed in [23], is needed in our system both to extend the operating range and also to ensure correct operation in the presence of IEEE 802.11a interference to compensate for the loss in SNR due to the addition of a bandpass filter in the receiver chain shown in Fig. 23. V. CONCLUSION A novel UWB positioning system has been presented. Experimental results of 5.24–6.37-mm real-time 3-D dynamic accuracy show its potential for millimeter accuracy even with a noncoherent architecture. Static 3-D accuracy of 0.73 mm show its potential for sub-millimeter accuracy when using a coherent system. This coherent architecture meets the 1–2-mm 3-D accuracy requirement for orthopedic surgical navigation systems. Experimental results in the time and frequency domains obtained in the OR were fit to the IEEE 802.15.4a channel model. This provides channel information to better design UWB indoor positioning systems to operate in the OR. The susceptibility of our UWB positioning system to 5-GHz WLAN interference was examined. This simulation highlighted the need for a high-performance bandpass filter at the front-end of our UWB receiver to mitigate the effects from 5-GHz interferers. Microcantilever and MEMS-based capacitive pressure sensors were presented. An overview of the performance of each of these sensors was given, as well as issues related to using the 315- and 433.92-MHz wireless telemetry bands for data transmission. The incorporation of UWB tracking with sensing elements provides a smart platform for future surgical navigation technologies not only with application to orthopedics, but also to neurosurgery, spinal surgery, and surgeries requiring an open MRI. Since electromagnetic tracking systems cannot operate in the presence of an MRI machine, and optical tracking systems, although they can be used, are cumbersome due to the LOS limitation, which is exacerbated when conducting a surgery and tracking with the constrictions of an open MRI, UWB positioning provides a powerful alternative to the standard tracking technologies when conducting surgery with an open MRI. ACKNOWLEDGMENT The authors would like to thank E. E. A. Fatah, University of Tennessee, Knoxville, for providing material related to smart preplanning, automated joint analysis, and post-operative anal-

[1] U. Varshney, “Using wireless technologies in healthcare,” Int. J. Mobile Commun., vol. 4, no. 3, pp. 354–368, 2006. [2] C. Pattichis et al., “Wireless telemedicine systems: An overview,” IEEE Antennas Propag. Mag., vol. 44, no. 2, pp. 143–153, 2002. [3] B. Evans, M. Mahfouz, and E. Pritchard, “Biocompatible MEMS electrode array for determination of three-dimensional strain,” in IEEE Int. Eng. Med. Biol. Soc. Conf., New York, Aug. 2006, pp. 4092–4095. [4] G. To, W. Qu, and M. Mahfouz, “ASIC design for wireless surgical MEMS device and instrumentation,” in IEEE Int. Eng. Med. Biol. Soc. Conf., New York, Aug. 2006, pp. 5892–5895. [5] E. Pritchard, M. Mahfouz, B. Evans, S. Eliza, and M. Haider, “Flexible capacitive sensors for high resolution pressure measurement,” in IEEE Sens. Conf., Lecce, Italy, Oct. 2008, pp. 1484–1487. [6] F. Langlotz, G. Zheng, and L.-P. Nolte, “Advanced technologies in navigation,” in Navigation and MIS in Orthopedic Surgery. Berlin, Germany: Springer, 2007, ch. 2, pt. VII, pp. 582–585. [7] J. Kowal, F. Langlotz, and L. Nolte, “Basics of computer-assisted orthopaedic surgery,” in Navigation and MIS in Orthopedic Surgery. Berlin, Germany: Springer, 2007, ch. 1, pt. I, pp. 2–8. [8] “Polaris Spectra & Polaris Vicra technical specifications,” Northern Digital Inc., Waterloo, ON, Canada, 2008. [Online]. Available: http:// www.ndigital.com/medical/polarisfamily-techspecs.php [9] “3D guidance medSAFE,” Ascension Technol. Corporation, Milton, VT, 2008. [Online]. Available: http://www.ascension-tech.com/docs/ 3DGuide%20medSAFE.pdf [10] B. Merkl, M. Kuhn, M. Mahfouz, and D. DeBoer, “Surgical navigation systems: Evaluating electromagnetic versus optical technology in the OR,” presented at the Sci. Exhibit Amer. Acad. Orthopaedic Surgeons Annu. Meeting San Francisco, CA, Mar. 2008. [11] D. Clarke and A. Park, “Active-RFID system accuracy and its implications for clinical applications,” in IEEE Int. Comput.-Based Med. Syst. Symp., Salt Lake City, UT, Jun. 2006, pp. 21–26. [12] “Hardware datasheet,” Ubisense, Cambridge, U.K., 2007. [Online]. Available: http://www.ubisense.net/media/pdf/ Ubisense%20System%20Overview%20V1.1.pdf [13] “Sapphire DART (RTLS) product data sheet,” Multispectral Solutions Inc., Germantown, MD, 2008. [Online]. Available: http://www.multispectral.com/pdf/Sapphire_DART.pdf [14] S. Roehr, P. Gulden, and M. Vossiek, “Method for high precision clock synchronization in wireless systems with application to radio navigation,” in IEEE Radio Wireless Symp., Long Beach, CA, Jan. 2007, pp. 551–554. [15] M. Vossiek, A. Urban, S. Max, and P. Gulden, “Inverse synthetic aperture secondary radar concept for precise wireless positioning,” IEEE Trans. Microw. Theory Tech., vol. 55, no. 11, pp. 2447–2453, Nov. 2007. [16] B. Waldmann, R. Weigel, P. Gulden, and M. Vossiek, “Pulsed frequency modulation techniques for high-precision ultra wideband ranging and positioning,” in IEEE Int. Ultra-Wideband Conf., Hannover, Germany, Sep. 2008, pp. 133–136. [17] “LPR-2D datasheet,” Symeo GmbH, Neubiberg, Germany, 2008. [Online]. Available: http://www.symeo.com/cms/upload/PDF/ Datasheet_LPR-2D.pdf [18] C. Meier, A. Terzis, and S. Lindenmeier, “A robust 3D high precision radio location system,” in IEEE MTT-S Int. Microw. Symp. Dig., Honolulu, HI, Jun. 2007, pp. 397–400. [19] M. Mahfouz, C. Zhang, B. Merkl, M. Kuhn, and A. Fathy, “Investigation of high accuracy indoor 3-D positioning using UWB technology,” IEEE Trans. Microw. Theory Tech., vol. 56, no. 6, pp. 1316–1330, Jun. 2008.


[20] C. Zhang, M. Kuhn, B. Merkl, M. Mahfouz, and A. Fathy, “Development of an UWB indoor 3-D positioning radar with millimeter accuracy,” in IEEE MTT-S Int. Microw. Symp., San Francisco, CA, Jun. 2006, pp. 106–109. [21] C. Zhang, A. Fathy, and M. Mahfouz, “Performance enhancement of a sub-sampling circuit for ultra-wideband signal processing,” IEEE Microw. Wireless Compon. Lett., vol. 17, no. 12, pp. 873–875, 2007. [22] M. Kuhn, C. Zhang, B. Mserkl, D. Yang, Y. Wang, M. Mahfouz, and A. Fathy, “High accuracy UWB localization in dense indoor environments,” in IEEE Int. Ultra-Wideband Conf., Hannover, Germany, Sep. 2008, vol. 2, pp. 129–132. [23] M. Kuhn, C. Zhang, S. Lin, M. Mahfouz, and A. Fathy, “A system level design approach to UWB localization,” in IEEE MTT-S Int. Microw. Symp. Dig., Boston, MA, Jun. 2009, pp. 1409–1412. [24] C. Zhang, M. Kuhn, M. Mahfouz, and A. Fathy, “Real-time noncoherent UWB positioning radar with millimeter range accuracy in a 3D indoor environment,” in IEEE MTT-S Int. Microw. Symp. Dig., Boston, MA, Jun. 2009, pp. 1413–1416. [25] M. Mahfouz, “Operating room of the future: Orthopedic perspective,” presented at the Cairo Int. Biomed. Eng. Conf. Invited Talk, Cairo, Egypt, Dec. 2008. [26] B. Merkl, “The future of the operating room: Surgical preplanning and navigation using high accuracy ultra-wideband positioning and advanced bone measurement,” Ph.D. dissertation, Mech. Aero. Biomed. Eng. Dept., Univ. Tennessee, Knoxville, TN, 2008. [27] A. Molisch, D. Cassioli, C. Chong, S. Emami, A. Fort, B. Kannan, J. Karedal, J. Kunisch, H. G. Schantz, K. Siwiak, and M. Z. Win, “A comprehensive standardized model for ultrawideband propagation channels,” IEEE Trans. Antennas Propag., vol. 54, no. 11, pp. 3151–3166, Nov. 2006. [28] A. Molisch et al., “IEEE 802.15.4a channel model—Final report,” IEEE, Piscataway, NJ, Tech. Rep., Doc. IEEE 802.1504-0062-02-004a, 2005. [29] L. Hentila, A. Taparungssanagorn, H. Viittala, and M. Hämäläinen, “Measurement and modelling of an UWB channel at hospital,” in IEEE Int. Ultra-Wideband Conf., Zurich, Switzerland, Sep. 2005, pp. 113–117. [30] M. Kuhn, M. Mahfouz, C. Zhang, B. Merkl, and A. Fathy, “Electromagnetic interference and its effects on ultra-wideband technology in the operating room,” in Int. Biomed. Eng. Conf., Cairo, Egypt, Dec. 2006, pp. 1–4. [31] S. Gabriel, R. W. Lau, and C. Gabriel, “The dielectric properties of biological tissues: III Parametric models for the dielectric spectrum of tissue,” Phys. Med. Biol., vol. 41, pp. 2271–2293, 1996. [32] A. V. Vorst, A. Rosen, and Y. Kotsuka, RF/Microwave Interaction With Biological Tissues. Hoboken, NJ: Wiley, 2006. [33] J. Collins, “RFID remedy for medical errors,” RFID J. May 28, 2004. [Online]. Available: http://www.rfidjournal.com/article/view/961/1/1 [34] N. F. Güler and U. Fidan, “Wireless transmission of ECG signal,” J. Med. Syst., vol. 30, no. 3, pp. 231–235, 2006. [35] “Path loss in remote keyless entry systems,” Maxim-IC Inc., Sunnyvale, CA, Appl. Note 3945, Dec. 15, 2006. [Online]. Available: http:// www.maxim-ic.com/appnotes.cfm/an_pk/3945 [36] B. Merkl and M. Mahfouz, “Unsupervised three-dimensional segmentation of medical images using an anatomical bone atlas,” in 12th Int. Biomed. Eng. Conf., Singapore, Dec. 2005, pp. 1–5. [37] M. Mahfouz, E. Abdel Fatah, H. E. Dakhakhni, R. Tadross, and R. Komistek, “Three-dimensional bone creation and landmarking using two still X-rays,” presented at the Annu. Amer. Acad. Orthopaedic Surgeons Meeting, San Francisco, CA, Mar. 2008, Podium 196. [38] M. Mahfouz, B. Merkl, E. Abdel Fatah, R. Booth, and J. Argenson, “Automatic methods for characterizing of sexual dimorphism of adult femora: Distal femur,” Comput. Methods Biomech. Biomed. Eng., vol. 10, no. 6, pp. 447–456, 2007. [39] M. Mahfouz, W. Hoff, R. Komistek, and D. Dennis, “A robust method for registration of three-dimensional knee implant models to two-dimensional fluoroscopy images,” IEEE Trans. Med. Imag., vol. 22, no. 12, pp. 1561–1574, Dec. 2003. [40] R. Wasielewski, D. Galat, and R. Komistek, “An intraoperative pressure-measuring device used in total knee arthroplasties and its kinematics correlations,” Clin. Orthopaedic Rel. Res., vol. 427, pp. 171–178, 2004.

2563

[41] W. Qu, S. Islam, G. To, and M. Mahfouz, “Micro-cantilever array pressure measurement system for biomedical instrumentation,” in IEEE Sens. Conf., Atlanta, GA, Oct. 2007, pp. 1009–1012. [42] B. Evans and M. Mahfouz, “Design optimization of a three-dimensional strain sensor using multiphysics finite element analysis,” in Int. Biomed. Eng. Conf., Cairo, Egypt, Dec. 2006, pp. 1–4. [43] J. Bellorado, S. S. Ghassemzadeh, L. J. Greenstein, T. Sveinsson1, and V. Tarokh1, “Coexistence of ultra-wideband systems with IEEE802.11a wireless LANs,” in IEEE Global Telecom. Conf., San Francisco, CA, Dec. 2003, pp. 410–414. [44] L. Zhu, H. Bu, and K. Wu, “Broadband and compact multi-pole microstrip bandpass filters using ground plane aperture technique,” Proc. Inst. Elect. Eng.—Microw., Antennas, Propag., vol. 149, no. 1, pp. 71–77, 2002.

Mohamed R. Mahfouz (S’98–M’01–SM’06) received the B.S.B.M.E. and M.S.B.M.E. degrees from Cairo University, Cairo, Egypt, in 1987 and 1992, respectively, the M.S.E.E. degree from the University of Denver, Denver, CO, in 1997, and the Ph.D. degree from the Colorado School of Mines, Golden, in 2002. From 1998 to 2002, he was Technical Director of the Rocky Mountain Musculoskeletal Research Laboratory, Denver, CO. In 2002, he became both Technical Director of the Center for Musculoskeletal Research and an Associate Professor with the University of Tennessee, Knoxville. His current research interests include medical applications of UWB, biomedical instrumentation, medical imaging, surgical navigation, MEMS bio-sensors, and 3-D bone and tissue reconstruction. He has authored numerous journal and conference papers and book chapters. He has received numerous National Institutes of Health (NIH) and National Science Foundation (NSF) grants.

Michael J. Kuhn (S’06) was born in Wheat Ridge, CO, in 1982. He received the B.S. degree in electrical engineering and B.S. degree in computer science from the Colorado School of Mines, Golden, in 2004, the M.S. degree in engineering science from the University of Tennessee, Knoxville, in 2008, and is currently working toward the Ph.D. degree in biomedical engineering at the University of Tennessee. Since 2005, he has been a Researcher with the Center for Musculoskeletal Research, University of Tennessee, Knoxville. He has authored/coauthored and presented at numerous international conferences in the fields of biomedical engineering and also microwave and antenna engineering. His current research interests include medical applications of UWB, numerical techniques in microwave engineering, and orthopedic surgical navigation. Mr. Kuhn was the recipient of the 2005 Ph.D. Fellowship presented by the College of Engineering, University of Tennessee.

Gary To (S’05) was born in Hong Kong, in 1982. He received the B.S. degree in biomedical engineering (with a minor in material science and engineering) from the University of Tennessee, Knoxville, in 2004, the M.S. degree in biomedical engineering from the University of Tennessee, Knoxville, in 2007, and is currently working toward the Ph.D. degree in biomedical engineering at the University of Tennessee. In Summer 2003, he was an Undergraduate Research Assistant with the Center for Musculoskeletal Research, Knoxville, TN. He has authored/coauthored and presented at numerous international conferences in the fields of MEMS and ASIC design for biomedical applications. His current research interests include bioinstrumentation and smart implant design. Mr. To has been a member of Tau Beta Pi since 2003.

2564


Aly E. Fathy (S’82–M’84–SM’92–F’04) received the B.S.E.E. degree, B.S. degree in pure and applied mathematics, and M.S.E.E. degree from Ain Shams University, Cairo, Egypt, in 1975, 1979, and 1980, respectively, and the Ph.D. degree from the Polytechnic Institute of New York, Brooklyn, in 1984. In February 1985, he joined the RCA Research Laboratory (currently the Sarnoff Corporation), Princeton, NJ, as a Member of the Technical Staff. In 2001, he became a Senior Member of the Technical Staff with the Sarnoff Corporation. With the Sarnoff Corporation, he was engaged in the research and development of various enabling technologies such as high-T superconductors, low-temperature co-fired ceramic (LTCC), and reconfigurable holographic antennas. He was also an Adjunct Professor with the Cooper Union School of Engineering, New York, NY. In

August 2003, he joined the University of Tennessee, Knoxville, as an Associate Professor. He has authored or coauthored numerous transaction and conference papers. He holds 11 U.S. patents. His current research interests include wireless reconfigurable antennas, see-through walls, UWB systems, and high-efficiency high-linearity combining of digital signals for base-station amplifiers. He has developed various microwave components/subsystems such as holographic reconfigurable antennas, radial combiners, direct broadcast antennas (DBSs), speed sensors, and LTCC packages for mixed-signal applications. Dr. Fathy is a member of Sigma Xi and Eta Kappa Nu. He was the recipient of five Sarnoff Outstanding Achievement Awards (1988, 1994, 1995, 1997, and 1999). He is an active member of the IEEE Microwave Theory and Techniques Society (IEEE MTT-S) International Microwave Symposium (IMS) Technical Program Committee, the IEEE Antenna and Propagation Symposium, and the IEEE Radio and Wireless Steering Committee. He is currently the Technical Program chair of the 2008 IEEE Radio and Wireless Conference.