This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and

Transcription

1 This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier s archiving and manuscript policies are encouraged to visit:

2 Sensors and Actuators A 162 (2010) Contents lists available at ScienceDirect Sensors and Actuators A: Physical journal homepage: Implementation of wireless body area networks for healthcare systems Mehmet R. Yuce The School of Electrical Engineering and Computer Science, University of Newcastle, University Drive, Callaghan, NSW 2300, Australia article info abstract Article history: Received 4 June 2009 Received in revised form 22 May 2010 Accepted 8 June 2010 Available online 12 June 2010 Keywords: Medical sensor network Body area network Telemetry Body sensor network Medical monitoring This work describes the implementation of a complete wireless body-area network (WBAN) system to deploy in medical environments. Issues related to hardware implementations, software and wireless protocol designs are addressed. In addition to reviewing and discussing the current attempts in wireless body area network technology, a WBAN system that has been designed for healthcare applications will be presented in detail herein. The wireless system in the WBAN uses medical bands to obtain physiological data from sensor nodes. The medical bands are selected to reduce the interference and thus increase the coexistence of sensor node devices with other network devices available at medical centers. The collected data is transferred to remote stations with a multi-hopping technique using the medical gateway wireless boards. The gateway nodes connect the sensor nodes to the local area network or the Internet. As such facilities are already available in medical centers; medical professions can access patients physiological signals anywhere in the medical center. The data can also be accessed outside the medical center as they will be made available online Elsevier B.V. All rights reserved. 1. Introduction As portable devices like cellular phones, pagers and MP3 players become popular; people start to carry such devices around their bodies. In 2001, Zimmerman [1] studied how such electronic devices operate on and near the human body. He used the term wireless personal area network (PAN). He characterized the human body and used it as a communication channel for intra-body communications. Later around 2001, the term PAN has been modified to body area network (BAN) to represent all the applications and communications on, in and near the body [2]. One of the most attractive applications to use BAN is in the medical environment to monitor physiological signals from patients. Wireless body-area network (WBAN) is a special purpose wireless-sensor network that incorporates different networks and wireless devices to enable remote monitoring for various environments [3,4]. One of the targeted applications of WBAN is in medical environments where conditions of a large number of patients are continuously being monitored in real-time. Wireless monitoring of physiological signals of a large number of patients is one of the current needs in order to deploy a complete wireless sensor network in healthcare system. Such an application presents some challenges in both software and hardware designs. Some of them are as follows: reliable communication by eliminating collisions of two sensor signals and interference from other external wireless Tel.: ; fax: address: mehmet.yuce@newcastle.edu.au. devices, low-cost, low power consumption, and providing flexibility to the patients [5,6]. A WBAN-based wireless medical sensor network system when implemented in medical centers has significant advantages over the traditional wired-based patient-data collection schemes by providing better rehabilitation and improved patient s quality of life. In addition a WBAN system has the potential to reduce the healthcare cost as well as the workload of medical professions, resulting in higher efficiency. There is already a number of monitoring systems developed or being used in medical centers [7 17]. The available medical monitoring systems are generally bulky and thus uncomfortable to be carried by patients. Most of the current effort has mainly been focused on the devices that are monitoring one or few physiological signals only. When multiple sensors are involved, wires are used to connect the sensors to a wearable wireless transmitter. Wired systems restrict patients mobility and comfort level, especially during sleep studies. Future implementation of medical monitoring necessitates the use of small, low-power sensor nodes with wireless capability [17 19]. Thus far there is no available standard for a wireless body-area network specifically targeting health care. Most popular wireless communication technologies and protocols proposed or used in medical monitoring systems are listed in Table 1. Existing monitoring systems use the short-range wireless systems such as ZigBee (IEEE ) [9 11], WLANs [5,8], GSM [12] and Bluetooth (IEEE ) [13 15]. To make the power consumption and the size of the device low, short-range devices like Bluetooth and ZigBee are mostly used with sensors to collect medical data from a patient body. Especially WLAN technologies are avoided for low power /$ see front matter 2010 Elsevier B.V. All rights reserved. doi: /j.sna

3 M.R. Yuce / Sensors and Actuators A 162 (2010) Table 1 Wireless technologies used in medical monitoring. MICS [6] WMTS [6] UWB IEEE ( ) [28] IEEE (ZigBee) IEEE (Bluetooth) WLANs (802.11b/g) Frequency band MHz , , MHz 3 10 GHz 2.4 GHz (868/915 MHz Eur./US) 2.4 GHz 2.4 GHz Bandwidth 3 MHz 6 MHz >500 MHz 5 MHz 1 MHz 20 MHz Data rate 19 or 76 kbps 76 kbps 850 kbps to 250 kbps 721 kbps >11 Mbps 20 Mbps (2.4 GHz) Multiple access CSMA/CA, polling CSMA/CA, polling Not defined. CSMA/CA FHSS/GFSK OFDMA, CSMA/CA Trans. power 16 dbm (25 W) 10 dbm and 41 dbm 0 dbm 4, 20 dbm 250 mw <1.8 db Range 0 10 m >100 m 1, 2 m 0 10 m 10, 100 m m sensor nodes because of their large size and power consumption used to provide longer ranges (i.e. 100 m). As these technologies may most probably be installed in medical environments due to other applications, medical gateway devices should be designed in a WBAN to interface with these wireless systems to provide a wireless link between the control unit and a mobile device (i.e. PALM) or between the control unit and Internet via a Wi Fi link. The low-data rate IEEE technology (ZigBee) has been the most popular short-range standard used recently in medical monitoring systems due to its low transmitter power [20,21]. Systems using Zigbee wireless platform may however suffer from the strong interference by WLANs which share the same spectrum and transmit at a larger signal power [22]. Installing an interference free medical network in a hospital may thus be quite challenging since there exist a lot of other wireless systems and equipments using the 2.4 GHz band. The device technologies operating at the 2.4 GHz ISM band should thus deal with the interference and coexistence issues when they are located in the same environment [23]. As can be seen in Table 1, in addition to unlicensed ISM bands, there are medical bands such as MICS (Medical Implant Communication Service) and WMTS (Wireless Medical Telemetry Service) that are specifically regulated for medical monitoring by communication commissions around the world [24 26]. The recent short-range, low-data rate, ultra-wideband (UWB) technology is another attractive technology that could be used for body-area network applications because of its regulated low transmitter power [7]. In addition to review and discuss the current attempts in bodyarea network technology, a WBAN system that has been designed for healthcare applications will be presented in this paper. 1 The proposed WBAN is a multi-hopping wireless medical network that uses the MICS band to obtain physiological data from sensors placed on or in the body and the WMTS band as an intermediate node for a longer wireless communication. The data is transferred to remote stations through the local area network or the Internet already available in medical centers as a part of their ICT (Information and communication technologies) infrastructure. Unlike the other medical sensor networks (they usually use 2.4 GHz ISM band); we use medical standards occupying the frequency bands that are mainly assigned to medical applications. The paper is organized as follows. Section 2 describes an overview of WBANs for the medical environment. It gives the concept of WBAN applications with their important design features. Section 3 presents a complete WBAN implementation. Different medical scenarios are defined for a real implementation in hospital environment. This section also discusses the hardware implementation details for the proposed prototype system. The multi-access protocol used for multi-sensor and multi-patient scenario is also given here. Section 4 presents computer programs used for record- 1 This work is an expanded version of the conference paper given in Ref. [6]. ing, monitoring and processing the medical data captured through the prototype system. An overall system performance evaluation as well as comparisons with the recent attempts of WBAN-based patient monitoring systems have been given in Section 5. And finally Section 6 concludes the paper. 2. Wireless body-area networks in medical environment The application of WBAN in a medical environment may consist of wearable and implantable sensor nodes that sense biological information from the human body and transmit over a short distance wirelessly to a control device worn on the body or placed at an accessible location. The sensor electronics should be miniaturized, low-power and detect medical signals such as electrocardiogram (ECG), photoplethysmogram (PPG), electroencephalography (EEG), pulse rate, pressure, and temperature. The collected data from the control devices are then transferred to remote destinations in a wireless body-area network for diagnostic and therapeutic purposes by incorporating another wireless network for long-range transmission (see Fig. 1). The monitoring devices currently used in medical centers are not completely wearable because their electronics are bulky and wires are used for connections to multiple sensors. Fig. 1(a) shows the current application of a sensor network used in some modern medical centers. A wireless control unit (i.e. CCU) is collecting information from sensors through wires and transmits to a remote station for monitoring. In this implementation, the control unit is cumbersome and using wires is not advised for the comfort of patients. As shown in Fig. 1(b) the future medical sensor network requires miniaturized and wearable sensor nodes that can communicate with the receiving device wirelessly. The system will consist of individual wireless sensor nodes that can transfer a person s physiological data such as heart rate, blood pressure, ECG via a wireless link, without the need of any wired connection. Each sensor will have wireless capability and its design will be optimized in terms of the physical characteristics of the physiological signal. Having individual wireless nodes is also very beneficial since not all patients require all the physiological parameters for diagnosis. In the future, sensor node electronics could be designed using flexible and stretchable technology so that sensor nodes can easily be embedded in textile (i.e. patient clothes) to improve the patient s comfort in a healthcare environment [27]. Low power operation and miniaturization are two essential physical requirements of sensor nodes as they determine the lifetime of the devices and their suitability to be worn by patients. Power consumption of sensor nodes is dominated by the operation of the wireless chip, RF (radio frequency) transmission and reception. Thus it is desirable to use a wireless platform that will provide low power consumption and has a minimum transmitted power while still meeting the required range of a body-area network.

4 118 M.R. Yuce / Sensors and Actuators A 162 (2010) Fig. 1. A wireless sensor network system detecting and transmitting signals from a human body: (a) current application of medical sensor network and (b) future application of medical sensor network targeted by wireless body area network. A study group of IEEE has been launched in November 2007 to work on the WBAN standardization [4]. One of the main tasks of the group is to investigate a secure wireless platform to be dedicated to a WBAN application only as the existing wireless standards are employed in many different applications and are not exactly optimized for healthcare systems. UWB is one of the wireless candidates considered by this new standard. One big advantage of UWB wireless technology is that its data rate ranges from 850 kbp to 20 Mbp which can be used for simultaneous monitoring of many continuous physiological signals such as ECC, EEG and EMG [7,28]. In addition, UWB wireless technology does not also present an EMI (electromagnetic interference) risk to other narrow band systems and medical equipments in healthcare since its transmitter power is low and the frequencies used are not crowded. Although the WBAN standard is still working on the development of transmission band for the body-area networks, the UWB wireless chips are not available commercially to apply in a WBAN at the moment. Although UWB claimed very low power initially in the literature, the attempts of such technology in the integrated circuits have exhibited power consumption more than that of the conventional narrowband short-range wireless chips [29]. High power consumption mainly comes from the design of the receiver as higher RF and analog gains are required at the front-end to receive the very low level transmit power. For an implant node in WBAN, UWB will cause high penetration loss because it operates at high frequencies (3 10 GHz for medical applications) [28]. This may significantly affect the performance and size of the implantable nodes. UWB receivers are already more complex than the narrowband system ones due the low transmitted signal. The additional loss from the skin will necessitate a higher gain at the front-end of the receiver. Providing high gain at high frequencies is not feasible in RF technology. One method to overcome this problem is to use a transmit-only (Tx-only) UWB sensor node [37]. In a transmit only WBAN system, the sensor nodes contain a transmitter without a receiver and should be programmed to send sensor data using a special multi-access protocol to reduce collisions in order to realize a multi-patient monitoring. It is envisioned that the sensor nodes of a WBAN will use a dedicated wireless link which will most likely be a narrow ISM band considering the current wireless technology developments. However the UWB technology could still be incorporated in a WBAN for the applications that require higher data rate as well as in case a coexistence issue arises. In the future, a WBAN system, in order to be applied in a medical environment, should incorporate the following significant design improvements and features [3,5,8 10,12]: Low-cost and low-size sensor node electronics design with wireless capability. A sensor node will be able to transfer date over a distance up to a few meters. And sensor nodes should be miniaturized so that they can be easily wearable. Energy optimization techniques should be developed by the combination of the link- and physical-layer functionalities of the wireless devices, leading to increase in the battery life and thus the lifetime of the sensor nodes. Sleep mode technique should be used so that the sensor nodes will spend most of their time in the low-power state when data transmission is not required. Physiological data should be categorized as crucial and noncrucial data for each patient. As an example, although it could be different with different patients, vital signals like ECG can be more crucial than temperature for certain patients. Thus a WBAN system should prioritize the crucial data. High gain miniaturized antennas should be developed for sensor nodes to increase the transmission reliability and to minimize interference hence reducing power consumption. Each sensor should be optimized according to its characteristics using a variable sampling rate. Unlike other sensor network systems, in a WBAN system, each sensor signal has a different frequency (i.e. not uniform) and thus each sensor node should be optimized according to its sensor s frequency band. In addition, considering that some non-crucial physiological signals like temperature information may only be measured for a longer period of time (e.g. every hour), a WBAN will result in better performances if an adaptive communication protocol is utilized to accommodate such differences in the system which will clearly make the system power efficient. In order to provide a long-range remote monitoring, several gateway devices should be developed to interface with the existing wireless systems in healthcare. These gateway devices will mainly be used to provide communications between the CCUs and the remote computers or mobile devices. Handover mechanism should be integrated in a WBAN that could be useful for free movement of patients in a large medical environment. It can be used to track patient throughout the hospital or can track patient locations when they are outdoor doing their daily activities. An alarm option may be included if a patient leaves a room or goes out of the range of a CCU. This can give the last location of patients that would allow staff to easily track patient location. Such features can be included in the system by detecting signals through several remote stations, which are usually known as soft handover. Another necessary component in a WBAN is the wirelessnetwork security. Key software components should be defined and developed to accommodate secure and effective wireless

5 M.R. Yuce / Sensors and Actuators A 162 (2010) Fig. 2. Targeted three different scenarios in our wireless body area network system for multi-patient monitoring in medical centers, (a) when the device is used individually, or (b) and (c) for multi-patient monitoring in a medical center, representing one room and one floor respectively. networking. Data should be only accessible by an authorized person at remote destinations. Hardware components and software programs should be coordinated together to provide secure and reliable communications. Lastly, a WBAN system should have its own standards for data collection and storing techniques, and also for the wireless links to eliminate coexistence issues. 3. A body-area network implementation for healthcare Currently we are developing a complete wireless body-area network that is based on different frequencies in order to eliminate interference issues as well as to apply to different environments. We use MICS, WMTS and 433 ISM bands to detect signals from the sensors on the body. The goal in a WBAN application is to dedicate one sensor node to one physiological signal as described in Fig. 1(b) to eliminate placing wires on the patient body. However, there maybe some clinical applications that requires the monitoring of more channels of the same physiological signal simultaneously to provide a good quality screening [43,44]. More channels may be required when considering EEG signal for brain activities. For the applications that require monitoring of more channels such as ECG/EEG/EMG, we incorporate the UWB technology to achieve a high data rate wireless link [28]. We also like to point out that we are working to interface our devices with IEEE (ZigBee) and Wi-Fi links to cover a large area of body-area network. The selection of wireless schemes for sensor nodes will depend very much on the environment that the sensor nodes will be used. The body-area network prototyping system presented in this paper uses a multi-hopping structure where the MICS band is used for gathering signals from sensors and WMTS is used to transmit the sensor data to remote stations allowing a longer range monitoring. These frequency bands are internationally available and are permitted for a remote monitoring of several patients simultaneously. The MICS band has a low emission power (25 W, comparable to UWB) leading to a lower power consumption, and will thus provide one of the most suitable transmission bands for medical sensor nodes [26,30]. Although a few incidents have been reported due to the interference from some local TV channels in USA, WMTS is still the most popular band for wireless telemetry used in hospitals [31]. Hardware electronics and software programs are developed for three scenarios in the proposed WBAN as shown in Fig. 2. The first scenario targets individual use in the medical center or can be used in home care. This wireless body-area network comprises of sensor nodes, a CCU that transmits data to a local PC and then to a receiver station (i.e. remote PC) at a medical center. After obtaining the physiological data from a human body, sensor nodes transmit those data to the CCU via the RF link using the MICS band. The CCU then re-packages the data and transmits to the local PC. The data collected at the local PC is transferred to a remote PC across the

6 120 M.R. Yuce / Sensors and Actuators A 162 (2010) Table 2 Physiological parameter ranges and signal frequencies. Parameter Range of parameter Signal frequency (Hz) ECG signal mv Respiratory rate 2 50 breaths/min Blood pressure (BP) mmhg 0 50 EEG V Body temperature C EMG (electromyogram) 10 V to 15 mv GSR (galvanic skin reflex) 30 V to 3 mv Fig. 3. A multi-hopping WBAN prototyping system for multi-patient monitoring. This prototyping system uses two medical standards: MICS for short distance and WMTS for long distance wireless transmission respectively. is used between the CCU and the base station allowing for a much longer range. This implies that the intermediate CCUs must be able to operate both at MICS and WMTS frequencies, providing a link between the nodes and the computer. The CCU remains in close range with the patient and may be attached to their belt for example when it is for individual use. The nominal wireless distance for the MICS link is around 10 m. The WMTS link targets a distance more than 100 m. network in a medical center or through Internet if the patient is at a different location than the medical center. In the second and third scenarios, more than one patient share a CCU box. The CCU can either be connected to a local PC in the room (Scenario-2) or it can transmit data wirelessly to a remote CCU box that is attached to a PC via another wireless link (e.g. WMTS (600 MHz)). In the latter case, the CCUs act as portable wireless gateway devices (e.g. intermediate devices) and hence the system will form a multi-hopping wireless networking. The arrangement in the last scenario can be used for one or more rooms in a medical center. A hardware setup to realize the Scenario-2 and -3 is given in Fig. 3. Patients physiological parameters are sent to the intermediate CCUs and then to the base station (a remote PC). The system consists of three networking structures. A network between sensors and CCUs, another network between CCUs and a base station, and the last communication is between base stations (This is mainly a LAN connection). Deploying a complete wireless medical system in a hospital environment requires monitoring of a few hundred patients. With this article we discuss issues related to the implementation of such a large-scale body-area network and introduce techniques that suit well for an implementation in hospitals. The existing and reported body-area network projects have mainly focused on the scenario given in Fig. 2(a) mostly with the wired sensors connections described in Fig. 1(a) [3,9 12,14]. We have particularly been interested in a WBAN system which deals with the implementations shown in Fig. 2(b) and (c), Scenarios 2 and 3 respectively. Two pieces of software are created in this project. The software residing at the local PC is named GATEWAY. The job of GATEWAY is to gather data from the CCU through RS232/USB 2 cable and forwards it to one or more remote PCs using TCP/IP sockets over an Ethernet network. The software residing at the remote PC is named as BSN application. The BSN application collates data from the local PC, interprets and stores them onto the remote PC to be analyzed later by health professionals. The base station (i.e. the remote PC) is capable of displaying all the received data on a graphical user interface (GUI) and is also capable of storing all the data in the database system of a medical center (see Section 4). The MICS regulations require that the output power of any terminal must be kept under 16 dbm and is not intended for long-range wireless connections [24,26]. To facilitate a long-range communication, a WMTS link operating in the MHz band 3.1. Sensor nodes and CCU hardware designs We develop our individual sensor nodes to detect and transmit the physiological signals listed in Table 2. Characteristics of these physiological signals are obtained from the public domain available on the Internet. Most physiological signals have low amplitudes and frequencies in nature, and occupy a small information bandwidth [43]. At such low frequencies and low amplitudes, some problems inherent to circuits need additional attention. For a reliable information transfer it is necessary that the interface electronics in the sensor nodes detect the physiological signals in the presence of noise and increase the signal-to-noise ration (SNR) of the detected signal for a better processing by the subsequent blocks of the sensor nodes. Sensor nodes are designed to be small and power efficient so that their battery can last for a long time. They collect the signals from a human body which are usually weak and coupled with noise. An amplification/filtering process is utilized first to increase the signal strength and to remove the unwanted signals and noise. Then an Analog to Digital conversion (ADC) stage is employed to convert the analog body signals into digital for a digital signal processing. The digitized signal is processed and stored in a microcontroller. The microcontroller will then pack the data and transmit over the air via a wireless transceiver. Fig. 4 shows the hardware implementation of our sensor nodes and the block diagram. In our system we designed sensor nodes that can measure up to four body signals for a single patient. One node is dedicated to one of continuous physiological signals such as ECC, EEG or EMG. Four sensor nodes electronics (i.e. 4-channel) are built on a common PCB board so that some electronics can be used interchangeably 2 We have used both wired USB and RS232 connections. USB has a faster data communication and removes the necessity of holding data in the base station. Fig. 4. An example of sensor node hardware designed.

7 M.R. Yuce / Sensors and Actuators A 162 (2010) Fig. 5. Schematic of analog front-end for detection of continuous physiological parameters. for testing purposes. Eventually each board will be used for only one sensor signal in a WBAN (explained in Fig. 1(b)). The antennas for this project are designed as a loop printed around the prototyping boards (see Fig. 4). An important feature in our design is that the signals in Table 2 will be grouped into critical and noncritical data during the transmission. The wireless protocol will be designed such that the priority will be given to the critical data. The detail of this discussion is presented in Section 3.2. Considering the signal bandwidths given in Table 2, except for the EMG signal, a sampling rate of around Hz will be necessary for the ADC in the microcontroller (the sampling rate should be a minimum of twice the highest frequency in the signal that is digitized). For an EMG signal, at least 1 khz is required in order to detect the high frequency section of EMG waveforms. The tradeoff between the reduction in sampling rate and the total power consumption of the ADC is determined by the choice of the specific physiological parameter used in the sensor node. In our sensor board, the PIC microcontroller is able to provide a sampling frequency up to 52 KHz. Thus we are able to set an optimum sampling frequency for each individual physiological signal. As shown in Table 2 the amplitude and frequency information of the continuous physiological signals are low. They require an amplifier with a relatively high gain and a low-pass filter with a low cut-off frequency. Thus the same analog front-end can be used to detect these continuous signals. The front-end of the sensor nodes (i.e. interface electronics) in our prototype uses an instrumentation amplifier (INA321) and an active low-pass filter (LTC6081), as shown in Fig. 5 [32]. Both components are selected because of their high common mode rejection ratio (CMRR) and low current consumption properties. INA321 has a CMRR of 94 db, while LTC6081 has a CMRR of 100 db. The circuit in Fig. 5 is running from a 3.3 V source with virtual ground set at 1.25 V. The shutdown pins are connected to the microcontroller. The active current consumption is 40 A and when operating in shutdown mode it consumes less than 1 A. Typically ECG and EEG signals have amplitude of less than 500 V with a usable frequency less than 100 Hz. The input signals are amplified by 60 db, INA321 adds 14 db, while LTC6081 provides a gain of 46 db with a cutoff frequency at 100 Hz. It is important to note that the required amplification and filtering can easily be adjusted for a specific signal by modifying the capacitors and the resistors in the circuit. As an example, in case of an EMG sensor node, the capacitor of the front-end should be modified to have a cut-off frequency of 500 Hz. To save power and space in the sensor nodes, the required notch filter to eliminate the 50/60 Hz DC noise is done at the remote PC via a software program. We designed two different CCU boards to realize the scenarios given in Fig. 2. One CCU is designed to be connected to a computer via the USB port (Fig. 6(a)). The other CCU (CCU-2) given in Fig. 6(b) functions as an intermediate device (i.e. wireless gateway) that presents the WMTS wireless link. Although both CCUs can be used for multiple patients monitoring, the first CCU type (CCU-1) can also be useful for private usage at patients home or for a single patient monitoring in a hospital. It can receive the physiological signals directly from sensors without using the gateway CCU. The hardware for CCUs requires a microcontroller and a wireless transceiver to coordinate all the activities, similar to the sensor boards. The CCU-1/sensor nodes consist of a transceiver (AMI52100 IC) from AMI semiconductor used for the MICS band generation (we also used CC1000 in some sensor nodes to generate 433 MHz ISM and WMTS bands for sensors-ccu wireless connections) and the microcontroller-pic16f887. In addition to these chips, we use another transceiver-the CC1010 chip from Chipcon (this chip contains CC1000 and a microcontroller built-in) on the intermediate CCU board (CCU-2) to obtain a wireless transmission and networking with the WMTS band. The CC1010 and CC1000 transceiver chips can be configured to transmit anywhere within 300 and 1000 MHz frequencies. They have been programmed to operate with one of WMTS bands in our prototype system. The wireless chips AMI52100 IC and CC1000 are selected in the project because of the following reasons: overall cost saving, low-power consumption, size, and the suitability of operating at the MICS, WMTS, and 433 MHz ISM bands. AMIS has a data rate capability of 19 kbps while Fig. 6. CCU boards. (a) CCU-1: central control unit connected to a computer via a serial cable USB, (b) CCU-2: Intermediate Central Control Unit (CCU). This device is shared by more than one patient and portable. It receives signals from patients via the MICS wireless link and transmits to a base station via the WTMS wireless link. It uses two transceiver chips (AMIS and CC1010) to provide both wireless links.

8 122 M.R. Yuce / Sensors and Actuators A 162 (2010) Table 3 Summary of devices used in sensor nodes and CCUs. Features Devices AMIS (52100) CC1000 CC1010 a PIC16F887 Size (mm) Modulation ASK/OOK FSK/OOK FSK Sensitivity 117 dbm 109 dbm 107 dbm Power Transmitter Tx: 25 ma Tx: 26.7 ma Tx: 26.6 ma <0.6 ma (active) Receiver Rx: 7.5 ma Rx: 7.4 ma Rx: 9.1 ma Data rate 19 kbps 76.8 kbps 76.9 kbps Sniffing b 500 na 200 na 0mA Memory (RAM) Additional features 10-bit, 22.7 KHz sampl. freq. 10-bit, 52 KHz sampling a CC1010 has a microcontroller built-in. b Sniff mode enables the receiver to wake up or operates at times to sniff received RF signals and then return to sleep or wait mode if a signal is not detected. CC1010 provides 76 kbps. Summary of features for the devices used are listed in Table 3. Our sensor board, which includes the transceivers, the microcontroller and the sensor front-end electronics, consumes a power of 90 mw for the transmit mode and less than 30 mw for the receive mode when operating with a supply voltage of 3.3 V. For our testing scheme given in Fig. 3, the ASK (amplitude-shift keying) is selected for the communication between sensors and CCU. And for the communication between CCUs and the based station, a frequency-shift keying (FSK) mode is selected. Two different modulation schemes are configured from AMIS and CC1010 wireless transceivers in order to avoid any possible interference effects between two wireless links. The antennas for both WMTS and MICS bands are designed with coils as shown in Fig. 6. All the boards are custom made except for the base station which is the CC1010 evaluation module obtained from Chipcon. Currently as a part of our wireless body-area network project, we design some CCU-2 devices in which the second wireless comprises of the standards such as ZigBee and the Wi-Fi in order to accommodate and interface with different wireless platforms. This will extend the application environment that would be useful for patients. Especially a wireless gateway device where the second wireless is a Wi-Fi link will allow some patients to visit different environments (libraries, work places and houses, etc.) where the device can connect with Internet wirelessly and thus will able to send data to the medical center for remote monitoring. Meanwhile the patient will have the opportunity to continue with his social activities Multi-access protocol for sensor nodes and CCU A number of medium access control (MAC) protocols have been proposed for medical sensor networks [17 21]. Main requirements of a MAC protocol are reliability, flexible transmission mechanism, high channel efficiency and a low end-to-end delay time. Mainly three classes of MAC protocols have been considered for wireless medical applications. They are TDMA (time division multiple access), polling and the contention-based protocols also known as the random access protocols [20,21,33]. The TDMA and polling protocols are entirely contention free but centralized in nature. The TDMA protocol introduces a strict synchronization requirement whereas a polling network introduces a high overhead of polling message transmission. TDMA and polling-based networks introduce a fixed delay due to use of the fixed frame structure and the cycle time respectively. The contention-based protocols such as ALOHA and CSMA (carrier sense multiple access) have been proposed for some sensor network applications. These protocols are distributed in nature and do not require any centralized control Fig. 7. Frame structures for polling protocol and CSMA/CA communication protocol. p is the time that a polling sensor wakes up before the polling signal arrives from the CCU. The polling frame includes information about the sensors (sensor-ids) that needs to be polled to get a data.

9 M.R. Yuce / Sensors and Actuators A 162 (2010) signal from the CCU. They are also dynamic in nature and offer minimum packet transfer delays when operating under low to moderate load conditions. The performance of a contention-based protocol could degrade when the total traffic load increases significantly, which is an unlikely scenario in a medical sensor network application. In addition to its carrier sense ability, CSMA can also use a collision avoidance (CA) mechanism (so called CSMA/CA) to allow only one sensor node to communicate and send a packet at one time to avoid the interference and collision between sensor nodes [20,21]. The CSMA/CA MAC protocol is used by the IEEE based Wi-Fi and the IEEE based WPAN (wireless personal area network) standards. The CSMA/CA protocol offers lower delay and reliable transmission of packets in small size networks like a WBNA-based medical network [16,35]. Thus the use of this mechanism is attractive for the transmission of critical data in medical centers. Unlike other sensor network applications, in a WBAN system, not all sensor signals are required to be monitored all the time. As an example, a temperature reading is usually taken every hour in hospitals and thus a schedule-based protocol would be more suitable for such sensor signals. Therefore we have decided to use the combination of a polling protocol and CSMA/CA MAC protocol to transmit the sensor data from multiple sensors to the CCU in our WBAN application. A critical data (e.g. ECG, it could change according to the status of a patient) will particularly be used in a CSMA/CA-based sensor node. A packet can be transmitted immediately without waiting for its turn eliminating any delay [34,35]. For non-critical data like temperature where the data collection is not necessary all the time, the polling network will be used for these nodes. In this case, the information will be sent whenever it is needed as it is not critical. Sensor nodes that are configured in a polling scheme will enter the CSMA/CA scheme during the data transmission so that it will not affect the transmission of the critical data. In our system we use a collision avoidance-based CSMA (CSMA/CA) by including the RTS/CTS (ready to send/clear to send) and ACK (Acknowledgement) signals to eliminate the collision probability (see Fig. 7) [16,21]. The sensor nodes in polling scheme will sleep most of the time and wake up for a duration required for a reliable data transfer. The polling sensor node is basically programmed to go sleep mode after a successful packet transmission as shown in Fig. 7(a). When data has been sent to the CCU successfully, then the CCU sends an acknowledgment of receiving data correctly (see ACK command in Fig. 7(b)). The sensor node will then go back to the sleep mode until a pre-defined time arranged by the microcontroller. It will wake up before a polling signal. If the polling signal from the CCU is requesting a transmission from a particular sensor node, that sensor node will then try to send data with the similar data format of CSMA/CA. It is assumed that the polling sensor nodes have the knowledge of the polling time the time that CCU will request data from a polling sensor node. This is a valid assumption for medical applications because medical data like temperature can have a pre-defined data reading times assigned by medical professions. The polling frame includes the information about all sensors (sensor-ids) that need to be polled to get a data. In other words, they are triggered to send data. The nominal active period of the polling nodes is 1 min (more than enough until a successful packet is sent in CSMA/CA mode) while the sleep mode time could be up to half an hour. Fig. 7(b) shows the details of the MAC protocol used to coordinate the transmission of all signals between sensors and the CCU-Computer. The MAC protocol uses the CSMA/CA packet transmission technique with the RTS/CTS messages. It is incorporated in the firm wares at both the sensor nodes and the CCU to provide a bi-directional communication, to control the wireless transmission and to prevent collision between sensor nodes. When a sensor node wants to transmit a packet to the station (i.e. CCU), first it checks the status of the transmission channel as shown in Fig. 7(b). If the channel is free i.e. no carrier signal is present, then the CCU waits for a certain time period and then transmits a packet if the channel remains free for a specified period as shown in the figure. If the CCU senses the channel busy then it backs off a random time and reschedules its transmission attempt at a later time. The RTS packet contains the ID of the transmitter (sensor) and a particular receiver (CCU) so that the transmitter and the receiver (CCU) can pair up for a packet transmission. After the RTS transmission, the sensor node moves into the wait state expecting a CTS packet from the CCU. The CCU will transmit a CTS signal after the short wait period known as the SIFS (short interframe spacing). Other sensor nodes in the network can read the RTS/CTS packet transmissions and will refrain from further packet transmission by calculating the channel occupancy period. This process can avoid collisions and can improve the packet transmission reliability. When a sensor node receives the CTS packet, it initiates a data packet transmission which would be followed by an ACK or a negative ACK (NACK) packet by the receiver station (CCU). The ACK or NACK responses are generated by the CCU after performing a packet error-check mechanism using the checksum field of the data packet. If the checksum fails, the CCU will ask for retransmission using the NACK packet else an ACK packet is transmitted. A packet might be corrupted either due to the transmission error caused by low SNR or due to a collision when two or more transmissions overlap in time. During a RTS/CTS packet transmission or a data packet transmission if any collision happens the sensor node either receives a NACK packet or it times out initiating a retransmission using the RTS/CTS procedures. The time out procedure allows a sensor node to come out of the wait state if the ACK or NACK packet goes missing due to transmission errors. Above discussions show that the CSMA/CA-based protocol can efficiently support transmission of physiological data packets in a medical networking environment. In case of a large traffic like in a hospital emergency ward, as described in Scenario 3 in Fig. 2, four or more patients will be assigned to a particular CCU unit. This way the number of sensor nodes assigned to a CCU is limited and thus the performance of the WBAN system will not be affected due to a large volume of people. When the patients are in action walking around, then each patient can be connected to one CCU as described earlier. In addition, using polling in a CSMA-based system will allow certain sensor nodes to stay silent for a certain amount of time, and thus only the important sensors data will be transmitted at a present time. This adaptive scheme also helps the WBAN to reduce the delay in the system or to increase the number of sensors to be monitored. In terms of power consumption, sensors nodes with polling mechanism have longer battery life comparing to CSMA/CA based nodes (the crucial signals). Sleep and wake up patterns can be controlled by the acquired signal and by the application. Wake up pattern could be periodic or it can be demand driven. Only a small portion of the microcontroller (oscillator and counters), which operates with a very low clock rate, always stays awake in the system and will trigger the awakening process. As a result of this process, significant power saving is achieved and hence battery life is increased. The power consumption of the polling based sensor nodes can be kept significantly low and thus their battery life can last several years [6,16]. In case of CSMA/CA communication, the nodes are always active either in transmission (the transmitter is active) or reception (the receiver is active). A CSMA node is also consuming power when its receiver is waiting to receive the ACK and the CTS packets. A continuously active CSMA-based node can last about three days, a lifetime similar to current cellular phones when using a battery of 300 mah.

10 124 M.R. Yuce / Sensors and Actuators A 162 (2010) Fig. 8. Live monitoring of multi-patients (physiological data presented in a graphical form at the remote PC). SensorID is used to define each patient. 4. Database, software programs and monitoring In order to monitor data, several computer programs have been developed during this project. The necessary software programs have been identified in Fig. 2. The software called GATEWAY is developed at the monitoring PC to control the communication with the CCU to get readings from sensors and then forward them through the network/internet to an application on a remote PC (at a medical profession center). While performing this task, the GATEWAY also verifies the data integrity and schedules retransmission if required. Another software program is developed at the remote PC (called BSN) that gets readings from GATEWAY via the network/internet. These readings are stored in the remote PC for analysis. The BSN application can collect and store readings automatically so that no person is required to be stationed at the application. It can undertake the administration of patients particulars such as assigning new sensor ID (i.e. userid) to patients, segregating sensor readings from different patients and storing them into the database. Using sensorid for each patient will ensure safety in healthcare environments when multiple patients are monitored. A graphical user interface (GUI) at the local PC as well as at remote PCs displays medical data. The GUI also allows the medical personnel to enter the patient s information. Both the data received from the CCU and the data sent to the BSN can also been shown by Fig. 9. A clean wireless ECG monitoring showing pulse rate as well. the GUI in text or graphical formats. The physiological signals of patients can be accessed by medical staff anywhere in a medical center as long as their computers are connected to the local area network in the building. An example of live monitoring from two patients scenario is shown in Fig. 8. It displays temperature and pulse rate information of two patients at the same time. Every sensor device has a unique Sensor ID and must be registered under a patient name before they are used [16]. In the event that an unregistered sensor node is used, all its readings received will be discarded by the BSN application. Although it is not implemented in the current prototype, a warning signal could be generated by the BSN application to warn the health profession to track the sensor node trying to send a signal. This feature will strengthen the reliability and safety in the implementation which would be useful for patients. The monitoring of the continuous signal like ECG, EMG and EEG is more complicated compared to parameters such as pulse rate and temperature signals. Unlike temperature and pulse rate, the information sent to the CCU requires a continual and undisturbed sampling period as high as 400 samples per second. Each sample will generate a data rate of 4000 with a sampling size of 10 bits. To allow for MAC overhead and packet retransmission, the baud rate (i.e. data rate) for the RF link should be at least twice this rate. The transceivers used in this prototype are able to transmit such data rates. Fig. 9 is an ECG signal obtained from our set up. Each sensor node representing only one patient can only have one ECG. In order to eliminate the DC noise (50 Hz/60 Hz interference) a recursive filter taken from [36] has been software implemented to obtain an accurate ECG signal. Shown in Fig. 10, by clicking on 50 Hz filter, the recursive notch filter operates on the received ECG signal. A database server has been developed to maintain data integrity which is necessary for big medical centers. A monitoring of ECG for an individual particular is shown in the Fig. 10. In the GUI, a more detailed of patients particulars can be seen from the database. Clicking on a patient shows the sensors attached to them and their personal information/picture. Clicking on a sensor displays the sensor s information (e.g. interval). It also depicts the number of recordings that are available. Single clicking a record shows when the record was made. Double clicking displays the record on the graph. In the following figure (Fig. 10), it is arranged for an ECG

11 M.R. Yuce / Sensors and Actuators A 162 (2010) Fig. 10. An example of live monitoring of ECG via database. Note that ECG electrodes were reversely polarized during measurement taken. monitoring. It is very handy to have a data file compatible with Matlab program for signal analysis as mentioned earlier. 5. Performance evaluation and discussion One way to analyze the performance of a WBAN system is to measure the end-to-end delay during monitoring of patients. The end-to-end delay performance herein refers to the time taken for a packet to be transmitted from the sensor nodes to the monitoring computer. In order to evaluate the performance of the proposed WBAN, we have analysed the data received from an active patient (subject) when moving from a CCU towards other patients, as shown in Fig. 11. To observe the end-to-end delay of sensor readings from this patient body, we checked the arrival times of the Fig. 11. A block diagram of real-time multi-patient monitoring to measure the performance of the WBAN system. Fig. 12. Monitoring screen: monitoring, text data at the remote PC, estimating data arrival times from stored data in database.

12 126 M.R. Yuce / Sensors and Actuators A 162 (2010) Fig. 13. Data arrival times at the remote computer. stored data (i.e. the received data) at the remote PC. Fig. 12 shows an example of our monitoring signals and the stored data (as a text) where the arrival times can be observed. Difference between two readings in time will be the end-to-end delay performance of the system. At some point, we could observe a couple of different data readings in the same second, which indicates that the delay performance is less than one second. We have calculated the value of the delay simply dividing one second (i.e. 1 s) by the number of data readings seen in a second for such occasions. In our test environment, we have measured the delay performances for distances of 1.5, 3, 5, 10 and 15 m. At these distances, the active subject was very close to other patients, that is where the maximum collision between sensor nodes could happen. The monitoring and the recording of the proposed WBAN system are done when the Wi-Fi link and a few Bluetooth devices are operating in the environment. These devices will not affect our prototype because the transmission frequency of our prototype is different. Selecting medical bands for a WBAN application is one of the advantages to eliminate external interferences. The measurement is taken in a room 15 m long and 4 m wide. We have implemented error-checking code in our CCU where only a correct packet is accepted. An erroneous packet is discarded and resubmission is requested which is also part of the CSMA/CA protocol explained in Section 3.2. In a CSMA/CA protocol, a sensor node sends an RTS signal to the CCU. The RTS signal includes the ID of the sensor node. Once the CCU sends a CTS signal back to that particular sensor node, a communication link is then established. Other sensor nodes in the environment will also read these RTS/CTS signals and will back off for a period of time 3 if the channel is not free. We also used a packet error-check mechanism (using the checksum field of the data packet) for the communication protocol between the gateway CCU and the local PC. If the checksum fails, the PC will ask for a retransmission using the NACK (negative acknowledgement) packet else an ACK packet is transmitted for a successful data receiving process, in the similar way done in CSMA/CA. The medical signal in our WBAN system will not be recorded or monitored unless it is a correct package. The packet error rate how- ever increases the number of retransmission and thus will increase the end-to-end delay to display correct medical data [35]. A packet might be corrupted either due to the transmission error caused by a low SNR (a parameter related to the distance) or due to a collision when two or more transmissions overlap in time. When the active subject moves close to other subjects (Fig. 11), the packet from the desired subject will be corrupted at a higher rate due to the transmission channel condition which requires retransmissions if there is overlap between two patient signals. Especially these events occur frequently when the sensor node is at a far distance away from the CCU. This has particularly been observed at the distance of 15 m. At this distance, we were not able to see any data being recorded, when the active sensor node is blocked by the active patient body (when the patient turns his back to the CCU). This is expected because at this distance, the received data power level is less than the receiver s sensitivity given in Table 3. In our test environment, when the active subject is around other subjects, he has been asked to do few routine actions such as sitting, standing, facing, and turning back with the same amount of time at the distance of 1, 1.5, 3, 5, 10 and 15 m. This routine action has been repeated at least ten times. And the average and worst case delays have been recorded and plotted in Fig. 13. The worst case delay was observed as 20 s when the distance between the active subject and the gateway CCU is 15 m. A medical data can be monitored less than 300 ms when the active subject is within distance of 1m and facing the CCU. When the active patient is doing all sort of activities e.g. sitting standing, turning around the other patients, the average delay is 1 s up to 5 m and 2 s at the distance of 10 m. These values show that our WBAN system exhibits a correct, timely and reliable communication performance up to 10 m with a time performance less than 2 s for a multi-patient scenario. We have observed that under the worst condition the average retransmission rate is 5 for a distance up to 10 m. The distance of 10 m has been our initial target during the development of the boards. Such a distance is sufficient for multi-patient monitoring. Usually the CCU is targeted to be placed within 2 m of the patients. However having 10 m will provide patient more flexibility and free movements in hospital environment Comparison of WBAN systems 3 During these backoff periods, the sensor nodes can operate in sniff mode (see Table 3) to reduce power consumption. In sniff mode, the sensor node does not send any signal, it is only in receive mode, by trying to receive a simple signal like ACK or NACK to follow the channel. This section discusses and compares the existing WBAN systems in the literature in terms of categories given in Figs. 1 and 2. The goal here is to compare these existing systems in terms their

13 M.R. Yuce / Sensors and Actuators A 162 (2010) Table 4 Comparisons of current implemented WBAN systems from literature. WBAN systems Wireless device Sensor Category a Comments Jovanov et al. [3] ZigBee (2.4 GHz) Activity sensor Similar to Fig. 2(a) The control device connects with a computer for data display directly or wirelessly using a Wi-Fi link. It is a multi-hopping system with multi-sensors; however one patient data has been implemented. No internet transmission has been indicated Gao et al. [9] IEEE (MICAZ) Pulse oximeter, blood pressure Anliker et al. [12] GSM Blood pressure, SpO2, one lead ECG Park et al. [17] nrf24e1 radio (2.4 GHz) Fig. 2(a) Figs. 1(a) and 2(a) Sensor nodes directly communicate with a tabled PC. A secure web portal to allow sharing real-time patient information (Internet connection) Multiple sensors have been integrated in one handheld device ECG Fig. 2(a) One of the smallest custom-made sensor nodes in the literature. The ECG sensor communicates to a computer wirelessly or a base station with Wi-Fi link. No multi-patient scenario Espina et al. [38] IEEE ECG, PPG, blood Fig. 2(a) Single user data is presented (2.4 GHz) pressure O Donovan et al. [42] ZigBee Motion, blood pressure, ECG Fig. 2(a) Data is sent to a base station (computer) wirelessly. No multi-patient data is presented Chen et al. [41] IEEE (ZigBee) 8-Channel EEG Figs. 1(a) and 2(a) Single patient. Internet data transmission is provided. No wireless gateways Zhang and Xiao [40] Bluetooth (2.4 GHz) ECG Fig. 2(a) and (b) Although it mentions it is done for two persons ECG, one continuous ECG monitoring data is shown. The control device is directly connected to a home computer via USB (no wireless gateways). No data transmission through Internet Huang et al. [39] ZigBee (2.4 GHz) No sensor Fig. 2 (a) Only one ZigBee device is used to communicate with an existing IEEE g device to measure interference in terms of BER. Does not provide monitoring signals Yuce et al. [28] UWB (4 GHz) 8-Channel ECG/EEG Fig. 2(a) The design targets monitoring of multi-channel continuous physiological parameters for implantable and wearable systems. No wireless gateways This work MICS/WMTS Pulse rate/temperature/ecg/eeg a Fig. 2(a), single body monitoring; Fig. 2(b) and (c), multiple body monitoring. Fig. 2(a) (c) The system has been designed to operate according to Fig. 1(b) and all the scenarios given in Fig. 2. Internet connection for long range suitability for a large-scale implementation in a medical environment. Comparisons of some of WBAN systems from literature are listed in Table 4. The column Category indicates whether the system demonstrates a multi-patient or a single patient monitoring. In Espina et al. [38] presents a IEEE based wearable bodyarea network that monitors continuous cuff-less blood pressure and ECG signals. The sensor data was measured on a single user. Date taken from a sensor is displayed on a PDA or a wristwatch device. In Ref. [9] sensor nodes directly talk to a tablet PC, a design similar to Fig. 2(a), and the stored data can then be seen by a secure web portal by medical professions. Anliker et al. [12] designed a wrist-worn device to measure multiple physiological parameters from one person. Sensors like blood pressure, SpO 2, one lead ECG all have been integrated in one handheld device. The data is then transmitted using a GSM data link to a computer similar to Fig. 1(a). The device acts the same as a regular cellular phone since it operates with the GSM network. However measurements more than one patient at the same time have not been provided. Another system in Ref. [17] designs very small wearable (probably one of smallest custom made sensor nodes) ECG sensors that communicate wirelessly with a base station connected to a computer. The system works similar to a multi-hopping if a Wi-Fi link is used. Similar to previous ones, this system also does not present any data for multi-patient measurements and monitoring. Data presented is from a scenario similar to Fig. 2(a). The available systems present medical monitoring mostly from a single body which is more suitable for homecare. Although it is one of the first and earlier WBAN implementations, E. Jovanov, 2005 in [3] has provided monitoring results for multi-sensors from a multihoping wireless link. To implement a WBAN system in a medical environment, the future WBAN systems should make sure that the implementation will operate the scenarios defined in Fig. 2(b) and (c). Real-time and simultaneous multi-sensor and multi-patient communication is required for future WBAN systems. In addition to real-time multi-patient monitoring capability, our system provides data transmission through the Internet allowing access to patient data at remote stations. It will be important for future WBAN systems to evaluate their design for multi-patient environment especially devices using a wireless platform at 2.5 GHz ISM band, due to the existing of the other wireless devices and equipments at this frequency. Such a validation will be crucial. Future WBAN systems should also develop small size sensor boards, as in Ref. [17], specifically targeted for WBAN applications to meet the size and power constraints. Most of the current system uses the commercial available ZigBee board which is designed for many other applications and thus is not entirely wearable. 6. Conclusions In this paper we present a multi-hopping network for a WBAN system that can be used in medical environments for remote monitoring of physiological parameters. The system is different than the existing implemented ones in that we consider monitoring of physiological signals from many patients simultaneously to represent a

14 128 M.R. Yuce / Sensors and Actuators A 162 (2010) real implementation in hospital environments. Issues related to a complete WBAN system to deploy in a medical center have been discussed. Portable and wireless gateway nodes are used to connect the sensor nodes to the local area network or the Internet already available in medical centers. This allows medical professions to access patient data anywhere in a medical center or even if they are away. Such a wireless body-area network system is very suitable to be used in hospitals environments to reduce human errors, to reduce health care cost, to provide more time to health professionals as well as to increase patients comfort level as they no longer need to be wakening up for periodic checks. A discussion of wireless technologies for use in a WBAN application has also been given in this paper. We are working to interface our MICS/WMTS devices with other wireless standards such as WLANs, ZigBee (IEEE ), mobile networks (e.g. GSM) or Bluetooth (IEEE ) in order to extend the applications in different environments. Such a heterogonous wireless network can help to coordinate collaborations across medical centers. Implementing a wireless body-area network will require integration of external and implanted nodes. Miniaturization of the sensor node electronics especially the sizes of the microcontroller, the wireless chip, the battery and low power consumption are current hardware related issues for small sensor nodes in a WBAN. The successful implementation of a complete WBAN system should operate and co-exists with other network device and should provide: wearable, wireless (no wire connections), easy to remove and attach sensor nodes, leading to increased mobility of patients and flexibility. Acknowledgements This work was supported in part by the Australian Research Council (ARC) under Discovery Projects Grant. I like to thank the following students Anthony Bott and Ng Peng Choong for their help in the developing boards. I also like to thank the anonymous reviewers for their comments and suggestions to strengthen the quality of the paper. References [1] T.G. Zimmerman, Personal Area networks: near-field intrabody communication, IBM Systems Journal 35 (3 & 4) (1996). [2] K.V. Dam, S. Pitchers, M. Barnard, From PAN to BAN: why body area networks? in: Proceedings of the Wireless World Research Forum (WWRF) Second Meeting, Nokia Research Centre, Helsinki, Finland, May 10 11, [3] E. Jovanov, A. Milenkovic, C. Otto, P. Groen, A wireless body area network of intelligent motion sensors for computer assisted physical rehabilitation Journal of Neuro Engineering and Rehabilitation 2 (6) (2005). [4] WBAN standard group. March, [5] A. Soomro, D. Cavalcanti, Opportunities and challenges in using WPAN and WLAN technologies in medical environments, IEEE Communications Magazine 45 (February (2)) (2007) [6] M.R. Yuce, C.K. Ho, Implementation of body area networks based on MICS/WMTS medical bands for healthcare systems, in: IEEE Engineering in Medicine and Biology Society Conference (IEEE EMBC08), August, 2008, pp [7] C.K. Ho, M.R. Yuce, Low data rate ultra wideband ECG monitoring system, in: IEEE Engineering in Medicine and Biology Society Conference (IEEE EMBC08), August, 2008, pp [8] Fischer, R., et al., SMART: scalable medical alert and response technology. [9] T. Gao, D. Greenspan, M. Welsh, R.R. Juang, A. Alm, Vital signs monitoring and patient tracking over a wireless network, in: IEEE-EMBS 27th Annual International Conference of the Engineering in Medicine and Biology, September, 2005, pp [10] C.A. Otto, E. Jovanov, E.A. Milenkovic, WBAN-based system for health monitoring at home, in: IEEE/EMBS International Summer School, Medical Devices and Biosensors, September, 2006, pp [11] C.H. Chan, C.C.Y. Poon, Wong, C.S. Raymond, Y.T. Zhang, A hybrid body sensor network for continuous and long-term measurement of arterial blood pressure, in: 4th IEEE/EMBS International Summer School and Symposium on Medical Devices and Biosensors, August, 2007, pp [12] U. Anliker, et al., AMON: a werable multiparameter medical monitoring and alert system, IEEE Transactions on Information Technology in Biomedicine 8 (2004) [13] C.M. Jong, R.S.H. Istepanian, A. Alesanco, H. Wang, Hardware design & compression issues in compact Bluetooth enabled wireless telecardiology system, in: IEEE 2nd International Conference on Broadband Networks, October, 2005, pp [14] M.F.A. Rasid, B. Woodward, Bluetooth telemedicine processor for multichannel biomedical signal transmission via mobile cellular networks, IEEE Transactions on Information Technology in Biomedicine 9 (2005) [15] J. Proulx, R. Clifford, S. Sorensen, D.J. Lee, J. Archibald, Development and evaluation of a Bluetooth EKG monitoring system, in: Proceedings of the 19th IEEE Symposium on Computer-Based Medical Systems, 2006, pp [16] M.R. Yuce, P.C. Ng, J.Y. Khan, Monitoring of physiological parameters from multiple patients using wireless sensor network, Journal of Medical Systems 32 (2008) [17] C. Park, P.H. Chou, Y. Bai, R. Matthews, A. Hibbs, An ultra-wearable, wireless, low power ECG monitoring system, in: Proceedings of IEEE BioCAS, 2006, pp [18] Wireless ECG patch by IMEC [19] EEG sensor by Icap [20] N.F. Timmons, W.G. Scanlon, Analysis of the performance of IEEE for medical sensor body area networking, in: Sensor and Ad Hoc Communications and Networks (IEEE SECON), 2004, pp [21] N. Golmie, D. Cypher, O. Rebala, Performance analysis of low rate wireless technologies for medical applications, Computer Communications 28 (2005) [22] I. Howitt, J.A. Gutierrez, IEEE low rate wireless personal area network coexistence issues, in: IEEE Wireless Communications and Networking Conference (WCNC), 2003, pp [23] A. Sikora, V.F. Groza, Coexistence of IEEE with other systems in the 2.4 GHz-ISM-Band, in: Proceedings of IEEE Instrumentation and Measurement, May, 2005, pp [24] FCC Rules and Regulations, MICS Band Plan, Table of Frequency Allocations, Part 95, January [25] Wireless Medical Telemetry, FCC Part 95, [26] S. Hanna, Regulations and Standards for Wireless Medical Applications, ISMICT, [27] Carta, et al., Design and implementation of advanced systems in a flexiblestretchable technology for biomedical applications, Sensors and Actuators A 156 (2009) [28] M.R. Yuce, H.C. Keong, M. Chae, Wideband communication for implantable and wearable systems, IEEE Transactions on Microwave Theory and Techniques 57 (October (Part 2)) (2009) [29] R. Gharpurey, P. Kinget, Ultra Wideband: Circuits, Transceivers and Systems, Springer, US, [30] H. Higgins, In-body RF communication and the future of healthcare. Zarlink semiconductor. [31] R. Diefes, Medical telemetry then and now. medsun/audioconf files/. [32] November [33] S. Ullah, et al., On PHY and MAC performance in body sensor networks, EURASIP Journal on Wireless Communications and Networking, Volume 2009 (2009), Article ID [34] G. Bugler, M.R. Yuce, Communication protocols for a multi-hoping wireless body sensor network. Technical Document, University of Newcastle, November, [35] J.Y. Khan, M.R. Yuce, F. Karami, Performance evaluation of a wireless body area sensor network for remote patient monitoring, in: Proceedings of the 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, Vancouver, BC, 2008, pp [36] March, [37] Ho Chee Keong, M.R. Yuce, Analysis of a multi-access scheme and asynchronous transmit-only UWB for Wireless Body Area Networks, in: The 31st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 09), , September, [38] J. Espina, T. Falck, J. Muehlsteff, Y. Jin, M.A Adán, X. Aubert, Wearable body sensor network towards continuous cuff-less blood pressure monitoring, in: The 5th International Workshop on Wearable and Implantable Body Sensor Networks, June, 2008, pp [39] L. Huang, R. de Francisco, G. Dolmans, Channel measurement and modeling in medical environments, ISMICT, [40] Y. Zhang, H. Xiao, Bluetooth-based sensor networks for remotely monitoring the physiological signals of a patient, IEEE Transactions on Information Technology in Biomedicine 13 (November (6)) (2009) [41] H. Chen, W. Wu, J. Lee, A WBAN-based real-time electroencephalogram monitoring system: design and implementation, Journal of Medical Systems March (2009). [42] T. O Donovan, et al., A context aware wireless body area network (BAN), in: Proceedings of Pervasive Health Conference, [43] R.F. Yazicioglu, P. Merken, R. Puers, C. Van Hoof, 60 W 60 nv/ Hz readout front-end for portable biopotential acquisition systems, IEEE Journal of Solid- State Circuits 42 (May (5)) (2007) [44] A. Bastani, H. Kayyali, R.N. Schmidt, R. Qadir, P. Manthena, Wireless Brain Monitoring in the Emergency Department, in: IEEE-EMBS 27th Annual International Conference, 2005, pp

15 M.R. Yuce / Sensors and Actuators A 162 (2010) Biography Mehmet Rasit Yuce received the M.S. degree in Electrical and Computer Engineering from the University of Florida, Gainesville, Florida in 2001, and the Ph.D. degree in Electrical and Computer Engineering from North Carolina State University (NCSU), Raleigh, NC in December Currently he holds senior lecturer position in the School of Electrical Engineering and Computer Science, University of Newcastle, New South Wales, Australia. Between August 2001 and October 2004, he has served as a research assistant with the Department of Electrical and Computer Engineering at NCSU, Raleigh, NC. He was a post-doctoral researcher in the Electrical Engineering Department at the University of California at Santa Cruz in His research interests include wireless implantable telemetry, wireless body area network (WBAN), analog/digital mixed signal VLSI for wireless, biomedical, and RF applications. Dr. Yuce has published about 50 technical articles and received a NASA group achievement award in 2007 for developing an SOI transceiver. He is a senior member of IEEE.