Statistical Analysis of On-body Radio Propagation Channel for Body-centric Wireless Communications

Transcription

1 374 PIERS Proceedings, Stockholm, Sweden, Aug , 2013 Statistical Analysis of On-body Radio Propagation Channel for Body-centric Wireless Communications H. A. Rahim 1, F. Malek 1, N. Hisham 1, and M. F. A. Malek 2 1 Embedded, Networks and Advanced Computing Research Cluster (ENAC) School of Computer and Communication Engineering Universiti Malaysia Perlis, P. O. Box 77, d/a Pejabat Pos Besar, Kangar 01000, Malaysia 2 MISC LNG Japan MISC Berhad, Yokohama, Japan Abstract This paper presents the experimental investigation of on-body radio propagation channel utilizing textile monopole antenna at 2.45 GHz. The measurement campaign was carried out in the anechoic chamber, considering stationary movement of the human body. The position of the transmitted antenna was fixed on the right of the upper arm of the human body and the received antenna was varied on several different potential on-body locations for body-centric wireless communication (BCWC) applications. The investigation was aimed to characterize the reflection coefficient and path loss of on-body radio channel when the antenna was placed in the vicinity of human body. A statistical analysis of path loss was also performed. The results showed that the measured reflection coefficient of four on-body positions experienced an upward frequency shift at a minimum of 0.2% compared to the simulated results due to the body coupling effect. In the anechoic chamber, the highest path loss was found for right upper arm-left ankle link while the lowest path loss was observed on the right upper arm-right chest link, proving that the closer a received antenna to the transmitted antenna, the better signal reception will be obtained. Based on the measurement results, it could be seen that the lognormal distribution fits very well to the on-body radio channel for narrowband frequency. 1. INTRODUCTION BCWC has received a lot of attention recently [1 4]. Since BCWC is intended to be implemented on the user s body, it is more practical to utilize a textile antenna with an omni-directional pattern in BCWC applications as it can be integrated into clothing. On-body radio propagation channel has been extensively published in the open literature [1 4]. However, very few work reported the utilization of textile antenna for on-body radio channel [3]. Thus, this paper presents the experimental investigation of reflection coefficient and path loss characteristics using planar textile monopole antenna at 2.45 GHz. The stationary on-body propagation channel for different body positions is derived and statistically analyzed. 2. MEASUREMENT SETUP The experiment was carried out in an anechoic chamber at Electromagnetic Hyper Sensitivity (EHS) Laboratory to eliminate multipath reflections from surrounding environment. The antenna utilizes a portable Agilent Field Fox model number N9923A 2-port Vector Network Analyzer (VNA) to generate (transmit) and measure (receive) the signal. A total number of sampled points per acquisition N = 1001 is set. Measurement was performed on a female subject of weight 51 kg with a height of 1.49 m. Two planar textile monopole antennas were used in this measurement campaign [5, 6]. The transmitter antenna (Tx) was placed fixed at the right side of the upper arm (RU). The receiver (Rx) was placed on the 11 other positions: right chest (), left chest (LC), right waist (), left waist (), right thigh (), left thigh (LT), right ankle (), left ankle (), center of back (B), right back (RB), left back (LB) and left upper arm (LU). Fig. 1 shows the location of the Tx and positions of Rx antennas. Two 5 m low loss semi-rigid coaxial cables were used in the measurement campaign. The cables were wrapped with Eccosorb Flexible Broadband Urethane Absorber model: FGM-U-SA microwave absorbing foams to minimize the spurious radiation from, and coupling between, the coaxial cables. The measurement setup for on body shows in Fig. 1. A 10 mm separation was set between the antenna and the body. Table 1 shows the distance between the Tx-Rx for on-body measurement. Five sweep durations were performed for each location. The simulation was performed using CST HUGO body model software where the HUGO model was defined at mm 3 voxel resolution.

2 Progress In Electromagnetics Research Symposium Proceedings, Stockholm, Sweden, Aug. 12, Tx LC Received Antenna Transmitted Antenna LT Vector Network Analyzer Figure 1: On-body measurement setup. Placements of transmitted and received antennas. Measurement in anechoic chamber. Positions Distance (cm) Chest () Table 1: Distance between Tx-Rx for on-body measurement. Chest (LC) Waist () Waist () Thigh () Thigh (LT) Ankle () Ankle () Center (B) (RB) (LB) Upper Arm (LU) RESULTS AND ANALYSIS The measured reflection coefficient for on-body static at all positions is illustrated in Fig. 2. It is observed that the measured reflection coefficient of all on-body Rx placements shifted to the right up to 7.2% due to the body coupling effect. This result also shows that the textile monopole demonstrated reflection coefficient, S 11 < 10 db for all on-body locations. Voltage Standing Wave Ratio (VSWR) is a function of the reflection coefficient, which describes the power reflected from the antenna. Fig. 3 shows the VSWR for on-body static for free space and nine positions of textile monopole. From the graph, it is seen that VSWR for all positions and free space is less than 2. In general, if the VSWR is less than 2, the antenna matching is considered excellent. Fig. 4 shows the comparison between simulated and measured reflection coefficient for four on-body positions, i.e., LU, LB, and of textile monopole antenna. The result clearly showed that the minimum frequency detuning occurred when the antenna was placed on the left back by 0.2% as compared to the simulated reflection coefficient result. It is evident that the left back is the least affected 0 LC LT B RB LB LU Figure 2: reflection coefficient for several positions of Rx.

3 376 PIERS Proceedings, Stockholm, Sweden, Aug , VSWR Free Space VSWR LC LT B Figure 3: VSWR for stationary on-body in free spaces and positions of Rx (c) -22 Figure 4: Comparison between left upper arm, left back, (c) right chest, (d) right thigh for simulated and measured reflection coefficient. (d) location by the body coupling when textile monopole was used as Tx. However, this on-body location is impractical to be applied in BCWC as it will make the user feel uncomfortable if the antenna is to be attached to the clothing. Hence, left upper arm is chosen as Tx position since the frequency detuning is less than 2% compared to the simulation result. The path loss is defined as the ratio of received to transmitted power computed from the measured data, averaging over the measured frequency transfers at each frequency point [7]. Fig. 5 shows the path loss for 5 locations of stationary on-body. The result shows that highest path loss at 2.45 GHz was obtained for RU- link with a maximum value of 53 db due to the longest distance between Tx and Rx. Meanwhile the lowest path loss was observed on the RU- link with a maximum value of 43 db. Since the propagation distance is shorter between RU and

4 Progress In Electromagnetics Research Symposium Proceedings, Stockholm, Sweden, Aug. 12, On body Least Square fit (γ = 1.7) Path loss (db) Figure 5: Path loss of five Rx position for stationary on-body propagation. LU LB Path loss (db) log(d /d 0 ) Figure 6: and modeled path loss for stationary on-body channel. locations, the electromagnetic wave can propagate in a direct path from transmitter to the receiver. The on-body radio channel can be modelled as a linear function of the logarithmic distance d between transmitter and receiver, expressed as [2] ( ) d P L db (d) = P L db (d 0 ) + 10γ log + X σ (1) where PL db (d 0 ) is the average path loss at 0.1 m and γ is the path loss exponent. X σ represents a shadowing (large-scale) fading defined as variation of the local mean around the path loss, Gaussian distributed random variable with standard deviation σ in db. In order to obtain the average path loss at d 0 and the path loss exponent γ, a least square fit technique is applied. Fig. 6 shows the measured and modeled path loss value for stationary on-body radio propagation channel, involving nine on-body positions, i.e.,, LC,,,, LT,, and B. The path loss exponent for this case is γ = 1.7 and the mean path loss is 32.3 db. A shadowing factor is determined by computing the deviation between measured and the calculated average path losses. Fig. 7 presents the measured CDF of path loss for stationary on- body radio channel in the chamber fitted to normal distribution (σ = 2.6). This explains that human body shadowing plays insignificant role to the stationary on-body path loss variation when utilizing an omni-directional antenna. A statistical analysis is also performed to the measured path loss by fitting the data to an empirical distribution, lognormal distribution. The measured CDF of stationary on-body path loss is shown in Fig. 7. The result exhibits that the measured on-body path loss in the chamber is very well fit to lognormal distribution (µ = 3.82, σ = 0.10). A smaller spread of data, indicating by σ = 0.10, shows that there is a direct path of propagation occurred along the body surface. d 0 Figure 7: CDF of deviation from average path loss fitted to normal distribution, path loss in the chamber.

5 378 PIERS Proceedings, Stockholm, Sweden, Aug , CONCLUSIONS The stationary on-body radio channel involving different body positions was carried out in an anechoic chamber. The characteristics of the reflection coefficient and path loss of on-body radio propagation channel were studied. The channel model derivation and statistical analysis of stationary on-body radio channel were also performed. The measured path loss was fitted to an empirical distribution function. Results exhibit that the textile monopole obtained S 11 < 10 db for all on-body positions and demonstrated an upward frequency shift of four selected on-body locations at a minimum of 0.2% compared to the simulated results. The results also confirm the distance dependency between Tx-Rx of stationary on-body radio channel in non-reflecting environment. The measured path loss was very well fitted to the lognormal distribution. REFERENCES 1. Cotton, S. L., G. A. Conway, and W. G. Scanlon, A time-domain approach to the analysis and modeling of on-body propagation characteristics using synchronized measurements at 2.45 GHz, IEEE Trans. on Antennas and Propagation, Vol. 57, No. 4, , Apr Sani, A., Y. Zhao, Y. Hao, S.-L. Lee, and G.-Z. Yang, A subject-specific radio propagation study in wireless body area networks, 2009 Loughborough Antennas and Propagation Conference (PC), 80 83, Loughborough, UK, Nov , Michalopoulou, A., A. A. Alexandridis, K. Peppas, T. Zervos, F. Lazarakis, K. Dangakis, and D. I. Kaklamani, On-body channel modelling: Measurement and statistical analysis, 2010 Loughborough Antennas and Propagation Conference (PC), , Loughborough, UK, Nov. 8 9, Abbasi, Q. H., M. M. Khan, S. Liaqat, A. Alomainy, and Y. Hao, Experimental investigation of ultra wideband diversity techniques for on-body radio communications, Progress In Electromagnetics Research C, Vol. 34, , Rahim, H. A., F. Malek, I. Adam, S. Ahmad, N. B. Hashim, and P. S. Hall, Design and simulation of a wearable textile monopole antenna for body centric wireless communications, PIERS Proceedings, , Moscow, Russia, Aug , Rahim, H. A., F. Malek, I. Adam, S. Ahmad, N. B. Hashim, and P. S. Hall, On-body textile monopole antenna characterisation for body-centric wireless communications, PIERS Proceedings, , Moscow, Russia, Aug , Dabin, J. A., N. Ni, M. Haimovich, E. Niver, and H. Grebel, The effects of antenna directivity on path loss and multipath propagation in UWB indoor wireless channels, Proc. of IEEE Conf. Ultra Wideband Syst. Technol., , Newark, New Jersey, 2003.