Sensors and Actuators A 180 (2012) 137–147
Contents lists available at SciVerse ScienceDirect
Sensors and Actuators A: Physical journal homepage: www.elsevier.com/locate/sna
An ultra-wideband wireless body area network: Evaluation in static and dynamic channel conditions Chee Keong Ho a , Terence S.P. See b , Mehmet R. Yuce c,∗ a b c
School of Electrical Engineering and Computer Science, University of Newcastle, Callaghan, NSW 2308, Australia Institute for Infocomm Research, 1 Fusionopolis Way, #21-01 Connexis, Singapore 138632, Singapore Department of Electrical and Computer Systems Engineering, Monash University, Vic. 3800, Australia
a r t i c l e
i n f o
Article history: Received 14 December 2011 Received in revised form 26 March 2012 Accepted 29 March 2012 Available online 23 April 2012 Keywords: Wireless body area network Ultra wideband sensor network Medical sensor network
a b s t r a c t Wireless body area network is a collection of wearable wireless sensors placed around or in a human body that are used to monitor important information from a human body. A receiver (i.e. control unit) is required to connect these sensors to remote locations (i.e. hospital database and call centres). In this work an ultra-wideband (UWB) body sensor node has been designed and tested to analyze the realistic performance of a UWB-based wireless body area network. The results indicate that the locations of sensors and the control unit on a human body play an important role on the performance of the wireless body area network system. The work herein also investigates optimal receiver positions for different sensor configurations. The results are evaluated in both static and dynamic channel conditions based on data transmission from the UWB sensor node developed for wireless body area network applications. Four common sensor positions, namely the chest, head, wrist and waist and three receiver positions-chest, waist and arm are considered. The experiment is conducted in an Anechoic chamber to minimize the effects of the environment. In the static experiment, the subject under test remains motionless for the entire test duration. Under static channel conditions, it was seen that the transmission power can be reduced by 26 dB, when the receiver is positioned at the optimum point on the body. The evaluation of the dynamic channel condition is also performed by allowing the test subject to move the body as in a walking motion. Due to the body movements, the transmission power should be increased by 7 dB to maintain the same bit error rate as that of the static experiment. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Wireless body area network (WBAN) is a collection of wireless sensors placed around or in a human body that are used to collect important information wirelessly [1,2]. In recent years, there has been some interest in using ultra-wideband (UWB) wireless technology for wireless body area network (WBAN) applications [3,4] because of some important benefits it offers such as low-power transmitter, low radio frequency (RF) and electromagnetic interference (EMI) effects in medical environment, small size antenna [5], and high data rate [6]. UWB communication has been used for monitoring continuous medical signals [7,8], high speed medical monitoring such as electronic pills [9], as a wireless video system in hospital environment [10], and multi-channel neural recording to obtain a data rate as high as 100 Mbps for brain–computer interfaces [6]. UWB is also attractive for medical gait analysis
∗ Corresponding author. E-mail address:
[email protected] (M.R. Yuce). 0924-4247/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.sna.2012.03.046
and tracking due to the highly accurate performance that can be obtained [11]. In a WBAN system, very stringent requirements have been placed on the sensor node. The sensor nodes should be small, consume extremely low power and reliable. Therefore, it is very important to minimize the power consumption of a sensor node. For the current WBAN efforts in the literature, the commercially available narrowband wireless platforms such as Crossbow’s Mica nodes and Texas Instrument’ CC1010 and CC2400 platforms are used for the developments of WBAN sensor nodes. These nodes have been based on Bluetooth or ZigBee wireless modules [1]. Unlike narrowband WBAN nodes, the UWB chips/hardware platforms are not available in the commercial domain in order to use them extensively. It is thus very important to develop a wearable sensor node based on UWB communication to analyze communication around a human body. There are significant efforts in the literature in the design of UWB transceivers [3,12], however these transceivers should be low-power to utilize a small-size battery as well as integrated with other components of sensor nodes on a small electronic boards in order to make a UWB sensor system wearable for WBAN applications.
138
C.K. Ho et al. / Sensors and Actuators A 180 (2012) 137–147
A typical WBAN system consists of multiple sensor nodes and a single receiver node (i.e. control unit) attached to each user [1]. Although the locations of the sensor nodes are generally fixed, depending on the specific physiological signals to be measured, the receiver position (the controller unit), on the other hand, is not fixed. It is normally placed at a location identified by the system designer or at a location that enhances the user’s comfort level. Selecting an optimal position for the receiver can enable a reliable transmission with a lower transmission power. This will enhance the battery lifespan of the sensor nodes. Three popular positions, namely the waist, chest and arm that can be used for the location of a WBAN receiver are studied in this paper. The evaluation is performed with the test subject in both static and walking scenarios. Generally, there are two types of channel models for WBAN, static and dynamic channels. The channel models for WBAN have typically been obtained in the literature using a vector network analyser (VNA). For example in [13] the channel model for a static standing position is measured using a VNA. Another study in [14] is undertaken to obtain the channel model for a static seated position. The channel condition of a WBAN is always changing due to the movement of the body. Therefore a static channel model will not be a good representation of the network characteristics as it does not provide the complete operating scenario of a WBAN. The effect of body motions (i.e. dynamic channel) on the WBAN has been investigated by [15] using a VNA. The measurement is conducted by measuring the static channel condition for the different phases of walking motion. A real time measurement of the body motion is not technically feasible with a VNA because of its limited functions. Therefore, an alternative performance indicator is needed to evaluate the optimum receiver positions for both the static and dynamic channel conditions. In this paper, the effect of body movements on the performance of a UWB based WBAN is studied using the practical bit error rate (BER) performances that are obtained from a UWB hardware sensor platform specifically developed for an on-body usage. In addition to charactering the UWB propagation around the body, the BER performance indicator is also used to show that it is possible to arrange the locations of the receiver such that transmitted power levels of the sensor nodes can be controlled within the body area network. This paper is organized as follows: Section 2 describes the proposed UWB WBA system. In Section 3, the design specifications of the proposed UWB-WBAN system used in the experimental setup is presented. The optimum receiver positions for the static channel condition are investigated in Section 4. The effect of body movements on the different receiver positions is analyzed in Section 5. And the conclusion is given in Section 6. 2. UWB-WBAN system A generic definition of UWB is that the emitting signals have a fractional bandwidth larger than 0.2 or at least 500 MHz. The fractional bandwidth is defined as the bandwidth between the −10 dB lower and upper corner frequency. UWB is allowed to operate in 0–960 MHz and the 3.1–10.6 GHz band; however the effective isotopic radiated power must be kept below −41 dBm/MHz. 2.1. UWB sensor node The sensor node block diagram for wireless body area network applications is depicted in Fig. 1. The physiological signal from the electrodes enters the sensor node through the front-end amplification circuit and is processed by the microcontroller. The microcontroller then performs the analog to digital conversion, determines the transmission format, modulation scheme and sets
the data rate. The UWB pulses are generated by passing the narrow rectangular pulse trains through a band-pass filter (BPF), followed by an amplifier. The UWB pulse rate is independent of the data rate, therefore enabling the system to vary the number of UWB pulses per data bit [7]. The square narrow pulses from the pulse generator are multiplied using logic gates to form a modulated UWB signal. For an OOK (on–off keying) modulation, the multiplier is a simple AND gate. The band-pass filter has a bandwidth of 1 GHz centered at 4 GHz to ensure that the UWB spectrum mask requirements are met. A wideband antenna designed for 3–5 GHz with an antenna gain of 2 dBi is used for wireless transmission. Fig. 1(b) shows the photo of the UWB sensor node developed for the UWB-WBAN system. The sensor nodes are assembled on a four layers printed circuit board with dimensions of 27 mm (L) × 25 mm (W) × 15 mm (H) including the battery, which is sufficiently compact for use in a WBAN application. The node is able to produce narrow pulse ranging from 0.3 ns to 2 ns, with a variable pulse repetitive frequency of between 17 MHz and 170 MHz. After the BPF, the UWB pulse is amplified using a wideband low noise amplifier (LNA) to meet the regulated −41.3 dBm/MHz transmission power level. This amplifier has initially been included to guarantee that the amplitude of the UWB pulses is sufficient enough for the distance targeted by a WBAN application. The existence of this amplifier is also useful for the study presented in this paper as it helps to modify the power level of the transmitted signal. 2.2. Comparison of UWB systems Table 1 shows the detailed specifications of the existing wireless sensor platforms used for WBAN applications. Wireless sensor node technologies based on narrow bands especially 2.4 GHz ISM have been used in medical monitoring widely. The UWB node is compared to narrowband WBAN nodes because the existing UWB boards presented either in the literature or in the commercial domain are either not for a WBAN application or they have large physical dimension of electronic boards, which are not wearable. The presented UWB sensor node is superior to the existing narrowband ones in terms of power consumption, size and the high data rate capability [16]. The UWB-WBAN node is based transmit-only (UWB Tx-Only), which is utilized to avoid the use of power hungry receiver structure. The UWB-BAN nodes in this work accommodate different UWB pulse rate and asynchronous package transmission pattern to provide a multi-access communication [17]. There are basically two types of UWB systems: multi-band carrier based systems and impulse-radio (IR-UWB) systems. Multiband systems using Orthogonal Frequency Division Multiplexing (OFDM) modulation schemes with the spectrum divided into bands of 528 MHz is supported by WiMedia Alliance [18]. The WiMedia Alliance is a global organization certifying UWB devices for multimedia applications. The power consumption of multi-band systems is high due to complex real time signal processing required for the OFDM modulation. The average power consumption of a WiMedia compatible chip set is about 300 mW [19], therefore they are not suitable for battery powered devices like WBAN nodes (see Table 2). IR-UWB systems send very short pulses and simpler modulation schemes such as pulse position modulation (PPM) or OOK, which result in significant power savings. They are therefore more suitable for battery operated WBAN applications. Several IR-UWB transceivers have been proposed for WBAN applications recently. Table 2 provides a comparison table for the design of these UWB transceiver hardwares with our complete UWB sensor node. Unlike narrowband WBAN platforms most UWB based platforms in the literature have not been integrated with sensors, front-end circuits, microcontroller, receiver back-end processing units and graphical user interfaces (GUI) at the workstation. The unique feature of the UWB hardware platform in this paper is
C.K. Ho et al. / Sensors and Actuators A 180 (2012) 137–147
139
Fig. 1. (a) Block diagram of a UWB-WBAN sensor node, (b) a photo of the UWB sensor board.
Table 1 Comparison of some existing narrowband sensor nodes with the UWB-BAN node. Model
Company
Frequency
Data rate
Trans. power (dBm)
Physical dimension
Power consumption Tx Rx
Mica2 (MPR400) MicAz
Crossbow Crossbow
38.4 kbps 250 kbps
−24 to +5 −24 to 0
58 mm × 32 mm × 7 mm 58 mm × 32 mm × 7 mm
27 mA at 3.3 V 17.4 mA at 3.3 V
10 mA at 3.3 V 19.7 mA at 3.3 V
Mica2DOT Tmote Sky node T-node MICS Node [1]
Crossbow Moteiv SOWNet Monash University Monash University
868/916 MHz 2400–2483.5 MHz (IEEE 802.15.4) 868/916 and 433 MHz 2.4 GHz (IEEE 802.15.4) 868, 433, 915, or 315 MHz 402–405 MHz
38.4 kbps 250 kbps 52.2 kbps 76 kbps
−20 to +10 25–0 −20 to −5 −16
25 mm × 6 mm 66 mm × 32.6 mm × 7 mm Diameter of 23 mm 30 mm×75 mm
25 mA at 3.3 V 19.5 mAat 3 V 25 mA at 3 V 27 mA at 3.3 V
8 mA at 3.3 V 21.8 mA at 3 V 13 mA at 3 V 8 mA at 3.3 V
3.5–4.5 GHz
Up to 10 Mbps
−41 (dBm/MHz)
25 mm×27 mm
15 mA at 3 V
0 (no receiver)
UWB-BAN Node (this work)
the integration of the micro-controller that provides a multi-access communication in addition to controlling the pulse generation for an optimum signal transmission. The size of the sensor nodes developed is sufficiently small to be wearable and placed on a human body. Although the size of chips developed for UWB transmitters and receivers can be quite small in integrated circuit technology, they are required to be placed on an electronics board together with other components such as microcontroller, matching circuits, regulator, battery and antenna. 3. UWB-WBAN specifications and experimental setup The block diagram of the experiment setup is shown in Fig. 2. The UWB sensor node is attached to the body and transmits a
pseudorandom data sequence to the UWB receiver (e.g. control unit). The receiver receives the data, decodes it, and calculates the BER. As described above, the pulse generation technique in the UWB sensor node is based on direct modulation (i.e. no carrier is used). The signal entering the receiver passes through a receiver front-end circuit, which includes a 3.5–4.5 GHz band pass filter to eliminate the unwanted out-of-band signals, followed by an amplification of 48 dB using three wideband LNAs before down converting to the baseband signal using a mixer and a 4 GHz VCO (voltage controlled oscillator). The baseband signal passes through a low pass filter with 100 MHz bandwidth before going through the baseband amplification stage. The recovered UWB pulse is digitalized using a high speed ADC and processed by the FPGA before transferring the data to the laptop using a serial cable.
Table 2 Comparison of UWB sensor node with UWB hardware developments. Design
CPU
Frequency
Data rate
Power cons.
Size (mm)
Our Platform [3] (Only transmitter IC)
PIC18F14K22 N/A
10 Mbps –
10 nJ/bit 2 mW (50 pJ/pulse)
27 × 25 × 15 (sensor board) N/A
[20] [21] [22] [19] (Alereon)
N/A N/A N/A –
3.5–4.5 GHz 3.432 GHz, 3.960 GHz, and 4.488 GHz (500 MHz Bandwidth) 3.1–5.1 GHz 3–5 GHz 0.5–2.5 GHz 3.1–10.6 GHz
250 kbps–10 Mbps 10 Mbps 100 kbps 53–480 Mbps
– 0.35 nJ/bit 0.19 nJ/bit 300 mW (0.625 nJ/bit)
5 × 5 (chip size) ∼50 × 15 (board) 2.25 × 1.1 (chip size) 5 × 5 (chip size)
140
C.K. Ho et al. / Sensors and Actuators A 180 (2012) 137–147
Fig. 2. Block diagram of the experimental setup with UWB transmitter and receiver shown.
Bit error rates are obtained based on the experimental parameters in Table 3. As mentioned earlier, the UWB pulse rate is selected to be much higher than the actual data rate. This helps to increase the processing gain as well as the synchronisation process. Moreover the technique using multiple UWB pulses per data bit enables the sensor node to perform a gating operation. The Federal Communications Commission (FCC) has regulated the peak and average power of a gated UWB signal in 2005 [23]. By gating process, the average transmission power is reduced, while a higher peak transmission power is allowed. FCC has stringent requirements on UWB transmission, allowing a peak transmission power limit of 0 dBm and an average limit of −41.3 dBm/MHz [24]. The measurement of the average and peak power can be calculated easily using a spectrum analyser in practice. For the average power measurement, the resolution bandwidth is 1 MHz with an integration time of 1 ms. A resolution bandwidth of between 1 and 50 MHz can be used for the measurement of the peak power. The peak limit is dependent on the resolution bandwidth and varies according to: peak power = 20 log
resolution bandwidth 50
dBm
(1)
In our setup, the UWB transmitter is allowed to transmit at a maximum peak power level of −24.4 dBm based on a 3 MHz resolution bandwidth according to the gating operation utilised. A spectrum plot and time domain plot according to the parameters in Table 3 are presented in Fig. 3(a) and (b), respectively. As depicted, the average transmitted power used in our experiment is lower than the regulated power level of UWB communication.
Table 3 UWB transmission signal parameters used in the experiment. Parameter
Values
UWB pulse rate UWB pulse width UWB frequency band Data rate Modulation Maximum transmission peak power
25 MHz 2 ns 3.5–4.5 GHz 1 Mbps OOK −24.4 dBm
Fig. 4 shows an example of UWB pulses recovered wirelessly at the receiver before the analog-to-digital converter (ADC) (see Fig. 2). The measurement is taken with the transmitter placed 1 m away from the receiver. This recovered pulse is digitalized using a high-speed ADC and processed by the field programmable gate array (FPGA) before transferring the data to a laptop using a serial cable. The role of the FPGA is to process the received multiple UWB pulses and determine whether it is bit “1” or bit “0” before sending to the laptop. The BER is computed on the laptop using Matlab software package based on a known pseudorandom data sequence that is sent by the transmitter. The sensitivity of the proposed UWB receiver is −83.4 dBm. The sensor and receiver positions evaluated in this study are illustrated in Fig. 5. Most of the commonly monitored physiological signals are located on the front upper half of the body. Therefore the four sensor positions commonly used for physiological signal monitoring are chosen to evaluate the performance of the UWBWBAN system. Sensor 1 is placed at the ECG position, sensor 2 is at the EEG position, sensor 3 is on the wrist and finally sensor 4 is placed on the waist. For each sensor and receiver pair, five sets of readings are taken. Each set of readings consists of 10,000 data bits. It is known that the orientation of the antenna affects the measurement results [25]. As the goal of this on-body study is to identify the optimal receiver positions, all readings are taken with the antenna orientation that gives the strongest signal. Apart from the antenna orientation, the separation distance between the antenna and the body also affects the antenna’s performance. The performance degrades as the antenna is placed close to the body (less than 1 mm) due to mismatch in impedance, which increases the system losses [5]. To mitigate this problem, the sensor nodes and the receiver antenna are placed 10 mm away from the body in order to reduce the effect of the body on the antenna. The selection of 10 mm separation distance is a realistic choice because for any commercial products there will be a case to house the sensor nodes, and thus provide the separation distance required. The measurements are conducted in an Anechoic chamber to minimize the effects of the environment. The measurement setup used for off-body analysis is shown in Fig. 6, meanwhile the onbody measurement setup for an ECG position is given in Fig. 7. An omni-directional UWB antenna is used at the transmitter, and a
C.K. Ho et al. / Sensors and Actuators A 180 (2012) 137–147
141
Fig. 3. UWB signal generation from the UWB sensor node: (a) spectrum, (b) time domain plots.
Fig. 4. An example of recovered UWB pulse at the receiver.
directional UWB antenna is used at the receiver. This combination is selected as this provides the best performance [26]. The physical dimension of the UWB sensor node is 27 mm (L) × 25 mm (W) × 15 mm (H) with the battery attached which is small enough to be wearable (Fig. 7). 4. Optimal receiver positions for UWB-WBAN
Fig. 5. Sensors and receiver locations.
In this section, optimal receiver positions are investigated for four positions of main physiological signals, as shown in Fig. 5. Herein all readings are taken in a controlled environment where the subject under test remains static throughout each measurement. The comparison of performances is discussed based on a BER of 10−3 . A BER of 10−3 is selected as a performance indicator here in this work because it is a reasonable BER value for a sufficient data
142
C.K. Ho et al. / Sensors and Actuators A 180 (2012) 137–147
Fig. 6. Off-body measurement setup.
Fig. 7. An on-body measurement setup for ECG position.
recovery performance for the OOK modulation. It is important to note any other BER can easily be selected for our BER measurements to investigate performance of UWB sensor networks.1 The experimentally measured BER plots for the ECG sensor position are presented in Fig. 8. By using the peak transmission power given in Fig. 3, the UWB sensor node is able to operate reliably for all the receiver positions.2 It is observed that for ECG sensors, the receiver placed at the chest requires 12 dB less power compared to the receiver placed on the waist. The superior performance of the chest receiver with respect to the ECG sensor is mainly due to the proximity of the sensor to the receiver. The receiver placed at the arm exhibited the worst performance, requiring 17 dB more power compared to that of the receiver placed at the chest. It is obvious that the distance between the ECG sensor and the arm receiver is shorter compared to the waist receiver. However, the arm receiver is placed at the side of the body and thus does not have a direct line-of-sight with the ECG sensor. The measured BER plots for an EEG sensor position are presented in Fig. 9. The experimental results show that the best receiver position is on the chest, while the worst receiver position is on the waist. For a BER of 10−3 , there is only 8 dB difference between
1 It is important to note that the UWB receiver designed is able to detect the transmitted UWB pulses up to 2 m with a BER lower than 10−5 for most cases. In this section the transmitted power level is considered as it is important for the sensor node. Although the power level of the received signal could be obtained for each scenario, the transmitted power levels were used as a performance indicator for BER curves. BER plots are obtained wirelessly with a receiver having sensitivity of −83 dBm. 2 Note that the maximum transmission power in BER measurements is selected as −24 dBm because this was the maximum peak power level for the UWB sensor node configured for the targeted body area network application.
the best receiver and the worst receiver positions. The distance between the EEG sensor and the waist receiver is the furthest, thus requiring more gain. Although the arm receiver is closer to the EEG sensor, there is only a 3 dB gain when compared to the case of waist receiver. This is because the arm receiver is located at the right side of the body, while the EEG sensor and a waist receiver are positioned at the middle of the body. The BER plots for a wrist sensor are shown in Fig. 10. For the sensor on the wrist, the receiver placed at the arm performs much better than the receiver placed at the waist and chest. For a BER of 10−3 , the receiver placed at the arm requires 18 dB less power compared to the receiver placed at the waist. The receiver on the waist is partially blocked by the body; therefore the performance is greatly affected. Fig. 11 shows the BER plots when the UWB WBAN sensor is positioned at the waist. The sensor placed on the waist requires the lowest transmission power. A transmission power of −46 dBm is enough to operate with a BER of 10−3 for all receiver positions. For sensors on the waist, the best receiver position is on the waist, followed by the arm, and then the chest. The difference in the required transmit power for the best and worst receiver positions is only 7 dB. The low transmission power characteristic of the waist sensor makes it suitable for the gateway node to be further from the waist sensor and closer to other sensors that require higher transmission power. The performance of the proposed UWB sensor node with respect to three receiver positions is tabulated in Table 4. Different sensor node configurations are considered based on a BER of 10−3 . For a WBAN application, there may be more than one sensor active at one time. Thus it is important to observe the power saving when all or some of the main sensors are active. The best receiver position for each combination is highlighted. The best receiver position is evaluated based on the average transmit power required to attain a bit error rate of at least 10−3 for the all the sensor nodes in the particular configuration. From the results obtained, the chest receiver position is the most optimal receiver position whenever an ECG sensor is used, which is a predicted result. If the ECG sensor is not used and a wrist sensor is used, the most optimal receiver position in this case will be the arm. The receiver located at the waist is preferred when there is only one waist sensor in use. Differences shown under the column of the total transmit power in Table 4 are differences between the total transmit power required for an optimum receiver position and the worst receiver position. For example for the first row when all main sensors are used, difference = (Tecg + Teeg + Twrist + Twaist)waist − (Tecg + Teeg + Twrist + Twaist)chest = (−43 dBm + −31 dBm + −28 dBm + −53 dBm) − (−53 dBm + −41 dBm + −34 dBm + −46 dBm) = 19 dB, where Tecg, Teeg, Twrist, Twaist are the transmission powers for sensors ECG, EEG, Wrist and Waist, respectively. The transmit power can be reduced up to 26 dB by selecting the optimum receiver position. The 26 dB is calculated based on the total transmission power reduction for ECG, EEG and wrist sensors. There may be numerous configurations for the sensor nodes in a WBAN system, however this study considers the practical sensor node locations used in medical application. The receiver position at the middle of the chest is the preferred choice for most of cases especially when an ECG sensing is used.
5. Effect of body movements on the performance of WBAN The effect of body movements on the performance of the UWB WBAN is described in this section. The experimental setting and the test subject are similar to that of the static scenario. During testing, the subject under test walks normally with body swaying movements and arm-swinging motion. The regular body movements
C.K. Ho et al. / Sensors and Actuators A 180 (2012) 137–147
143
Fig. 8. BER performances for an ECG sensor position.
Fig. 9. BER plot for the EEG sensor position.
for the different sensors and receiver positions are highlighted in Table 5. The BER plots for ECG sensor based on the body movements are presented in Fig. 12. The required transmit power to achieve BER of 10−3 is almost identical to that of the static scenario. However, as the transmit power reduces, the BER performance for the receiver placed at the arm degrades much faster than those of the remaining two receiver positions.
Fig. 13 depicts the BER plots for the EEG sensor due to the body motions. The performance for all receiver locations degrades significantly compared to the ECG sensor scenario. The performance of the chest receiver is degraded by 5 dB, while the performances for the arm and waist receivers require 7 dB more transmission power. The higher transmission power requirement is caused by the head movement. The effect of head movements based on the signal strength measured from the oscilloscope is illustrated in
Fig. 10. BER plot for the wrist sensor position.
144
C.K. Ho et al. / Sensors and Actuators A 180 (2012) 137–147
Fig. 11. BER plot for the waist sensor position.
Table 6. The signal strength is measured from the receiver placed at the chest. When the head is looking up, the signal strength drops by 60%, and when the head is looking down, the signal strength is doubled. Therefore, It will be important for WBAN application to increase the transmission power of EEG sensor node to compensate for the signal loss due to head movements. The wrist sensor performance suffers most significantly due to body movements as illustrated in Fig. 14. Due to the arm swinging motion, which partially blocks the waist and chest receivers, both receivers are unable to attain the required BER within the transmission limit of the UWB sensor node. The
arm receiver is the only receiver position that the wrist sensor can operate reliably. The arm receiver performance is not included in this plot, as there was not any performance degradation resulted from the movements. The performance of the arm receiver during the movement is similar to that of the static scenario. In terms of the waist sensor location, there is no performance degradation for the waist and chest receiver positions caused by body movements. However, an arm receiver requires 7 dB more transmission power to attain the required BER of 10−3 as illustrated in Fig. 15.
Table 4 Optimum receiver positions for UWB-WBAN systems.
C.K. Ho et al. / Sensors and Actuators A 180 (2012) 137–147
145
Fig. 12. Effect of the body movements on an ECG sensor.
Fig. 13. Effect of body movements on an EEG sensor.
The optimal receiver positions for each sensor to operate effectively with body movements are presented in Table 7. The chest receiver position is still the preferred choice for different sensor configurations shown in Table 4. However, if a wrist sensor is used, the only possible receiver position is the arm. In summary, the WBAN performance degrades by a maximum of 7 dB due to the
effect of walking motion for all sensor positions evaluated in this study except for the wrist sensor. It is important to conclude that the reliability of a WBAN system depends largely on the receiver position. By selecting an appropriate receiver position, the effect of movements can be kept minimal for each individual sensor nodes.
Fig. 14. Effect of body movement on a wrist sensor.
146
C.K. Ho et al. / Sensors and Actuators A 180 (2012) 137–147
Fig. 15. Effect of body movement on the waist sensor.
Table 5 Simulated body movements for the different sensors and receiver positions. Sensor
Receiver locations
Simulated movements
ECG ECG ECG EEG EEG EEG Wrist Wrist Wrist Waist Waist Waist
Chest Waist Arm Chest Waist Arm Chest Waist Arm Chest Waist Arm
Body swaying movement from side to side Body swaying movement from side to side Arm swinging motion with body swaying movement Head turning motion with body swaying movement Head turning motion with body swaying movement Head turning motion with arm swinging motion Arm swinging motion with body swaying movement Arm swinging motion with body swaying movement Arm swinging motion with body swaying movement Body swaying movement from side to side Body swaying movement from side to side Arm swinging motion with body swaying movement
In summary, the chest area is the optimum receiver position in the static scenario for the different sensor node configurations, as long as an ECG sensor is involved. Even when body movement is considered, the chest area is still the preferred choice for the receiver position. However, if a wrist sensor is present in the network for the body movement scenario, the only possible receiver position is the arm. Thus it is important for an UWB WBAN implementation to configure body sensors and the control unit locations accordingly. Finally, this study shows that the performance of a UWB WBAN system can significantly be improved by carefully arranging the sensor node as well as the control unit locations on the human body. Acknowledgment This research was funded by Australian Research Council (ARC) under Discovery Projects Grant DP0772929.
Table 6 Effect of head motions on the received signal strength. Normal position
Head up
Head down
Head left
Head right
36 mV
12 mV
80 mV
20 mV
20 mV
Table 7 Optimum receiver positions for body movements. Sensors
Optimum receiver position
Performance
ECG EEG Wrist Waist
Chest Chest Arm Waist
1 dB loss 5 dB loss No performance penalty No performance penalty
6. Conclusion Ultra-wide band is an attractive wireless technology for medical monitoring systems and more importantly it does not present an EMI risk to other narrow band systems and medical equipments in healthcare environment since its transmitter power is quite low and the frequencies used are at very high frequencies (>3.5 GHz). In this paper, a complete UWB based WBAN node has been presented and its performance is analyzed using the practical bit error rate performances. Selecting an appropriate position for the receiver ensures that reliable transmission can take place with minimum transmission power levels. A lower transmission power will increase the battery lifespan of the WBAN sensor nodes. It is shown that by carefully selecting the receiver position, the transmission power can be reduced by up to 26 dB.
References [1] M.R. Yuce, Implementation of wireless body area networks for healthcare systems, Sensors & Actuators A: Physical 162 (July) (2010) 116–129. [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, 2001. [3] J.R. Ryckaert, et al., Ultra-wide band transmitter for low-power wireless body area networks: design and evaluation IEEE Transactions on Circuits and Systems I, Regular Papers 52 (December (12)) (2005) 2515. [4] M.R. Yuce, C.K. Ho, C. Moo Sung, Wideband communication for implantable and wearable systems, IEEE Transactions on Microwave Theory and Techniques 57 (2009) 2597–2604. [5] T.S.P. See, Z.N. Chen, X.M. Qing, Proximity effect of UWB antenna on human body, in: Asia Pacific Microwave Conference, December, 2009, pp. 2192–2195. [6] M. Chae, Z. Yang, M.R. Yuce, L. Hoang, W. Liu, A 128-channel 6 mW wireless neural recording IC with spike feature extraction and UWB transmitter, IEEE Transactions on Neural Systems & Rehabilitation Engineering 17 (August) (2009) 312–321. [7] H.C. Keong, M.R. Yuce, Low data rate ultra-wideband ECG monitoring system, in: The IEEE Engineering in Medicine and Biology Society Conference (IEEE EMBC08), August, 2008, pp. 3413–3416. [8] K. Takizawa, H.B. Li, K. Hamaguchi, R. Kohno, Wireless vital sign monitoring using ultra-wideband based personal area networks, in: Engineering in Medicine and Biology Society. EMBS. 29th Annual International Conference of the IEEE, 2007, pp. 1798–1801. [9] M.R. Yuce, T. Dissanayake, H.C. Keong, Wireless telemetry for electronic pill technology, in: Proc. IEEE Sensors, 2009, pp. 1433–2143. [10] E. Ayar, UWB Wireless Video Transmission Technology in Medical Applications, August, 2010. [11] El-Nasr M.A. Shaban, R.M. Buehrer, Toward a highly accurate ambulatory system for clinical gait analysis via UWB radios, IEEE Transactions on Information Technology in Biomedicine 14 (March) (2010) 284–291. [12] W. Rhee, N. Xu, B. Zhou, Z. Wang, Low power non-invasive UWB systems for WBAN and biomedical applications, 17–19 November, International
C.K. Ho et al. / Sensors and Actuators A 180 (2012) 137–147
[13]
[14]
[15]
[16] [17]
[18] [19] [20]
[21]
[22]
[23]
[24] [25]
[26]
Conference on Information and Communication Technology Convergence (ICTC) (2010) 35–40. K. Takizawa, T. Aoyagi, J.-i. Takada, N. Katayama, K. Yekeh, Y. Takehiko, K.R. Kohno, Channel models for wireless body area networks, in: Engineering in Medicine and Biology Society, 2008, EMBS 2008. 30th Annual International Conference of the IEEE, 20–25 August, 2008, pp. 1549–1552. C. Yifan, T. Jianqi, J.C.Y. Lai, E. Gunawan, L. Kay Soon, S. Cheong Boon, P.B. Rapajic, Cooperative communications in ultra-wideband wireless body area networks: channel modeling and system diversity analysis, IEEE Journal on Selected Areas in Communications 27 (2009) 5–16. A. Taparugssanagorn, C. Pomalaza-Raez, R. Tesi, M. Hamalainen, J. Iinatti, Effect of body motion and the type of antenna on the measured UWB channel characteristics in medical applications of wireless body area networks, in: IEEE International Conference on Ultra-Wideband, September, 2009, pp. 332–336. H.C. Keong, M.R. Yuce, UWB-WBAN sensor node design, in: The International Conference of the IEEE Engineering in Medicine and Biology Society, 2011. H.C. 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, 2009. http://www.wimedia.org/, 2012. http://www.alereon.com/products/chipsets/, 2012. T. Nakagawa, G. Ono, R. Fujiwara, T. Norimatsu, T. Terada, M. Miyazaki, Fully integrated UWB-IR CMOS transceiver for wireless body area networks, IEEE International Conference on Ultra-Wideband (September) (2009) 768–772. Yuan Gao, et al., Low-power ultra-wideband wireless telemetry transceiver for medical sensor applications, IEEE Transactions on Biomedical Engineering 58 (March (3)) (2011). R.K. Dokania, X.Y. Wang, S.G. Tallur, A.B. Apsel, A low power impulse radio design for body-area-networks, IEEE Transactions on Circuits and Systems—I 58 (July (7)) (2011). FCC 05-58: Petition for Waiver of the Part 15 UWB Regulations Filed by the Multi-band OFDM Alliance Special Interest Group, ET Docket 04-352, March 11, 2005. FCC 02-48 (UWB first report and order), 2002. Z. Yue Ping, L. Qiang, Performance of UWB impulse radio with planar monopoles over on-human-body propagation channel for wireless body area networks, IEEE Transactions on Antennas and Propagation 55 (2007) 2907–2914. T.S.P. See, T.M. Chiam, C.K. Ho, M.R. Yuce, Experimental study of optimal UWB antenna location for ECG application, in: Proc. the IEEE International Conference on Ultra-wideband (ICUWB2010), September 20–23, 2010.
147
Biographies Ho Chee Keong received the B.S. degree and Ph.D. degree in electrical engineering from University of Newcastle, Australia, in 2006 and 2011, respectively. From 2006 to 2007, he served as an electronic design engineer with ET Designers, specializing in LCD and touch panel design. In 2011, he joined the Institute of Microelectronics (IME), Singapore. His research Interests include UWB communication scheme, biomedical applications and system designs for wireless communications. Terence S.P. See received the B.Eng. and M.Eng. degrees in electrical engineering from the National University of Singapore in 2002 and 2004, respectively. In 2004, he joined the Institute for Infocomm Research, Singapore. He is currently holding the position of Senior Research Engineer in the Antenna Lab under the RF and Optical Department. His main research interests include antenna design and theory, particularly in small and broadband antennas and arrays, diversity antennas, antennas for portable devices, and antennas for on-body communications and bio-medical applications. 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 2004. He has served as a research assistant between August 2001 and October 2004 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 2005. He was a Senior Lecturer in the School of Electrical Engineering and Computer Science, University of Newcastle, New South Wales, Australia until July 2011. In July 2011, he joined the Department of Electrical and Computer Systems Engineering, Monash University, Australia. His research interests include wireless implantable telemetry, wireless body area network (WBAN), bio-sensors, MEMs sensors, integrated circuit technology dealing with digital, analog and radio frequency circuit designs for wireless, biomedical, and RF applications. Dr. Yuce has published more than 80 technical articles in the above areas and received a NASA group achievement award in 2007 for developing an SOI transceiver. He received a research excellence award in the Faculty of Engineering and Built Environment, University of Newcastle in 2010. He is an author of the book Wireless Body Area Networks published in 2011. He is a senior member of IEEE. He is also a member of the following professional societies of IEEE: IEEE Solid-State Circuit Society, IEEE Engineering in Medicine and Biology Society, and IEEE Circuits and Systems Society.