On the WBAN Radio Channel Modelling for Medical Applications Matti Hmlinen, Attaphongse Taparugssanagorn, Jari Iinatti Centre for Wireless Communications (CWC), University of Oulu PO BOX 4500, FI-90014 University of Oulu, Finland {matti.hamalainen, jari.iinatti}@ee.oulu.fi, attaphongset@gmail.com Abstract Radio channel modelling for wireless body area network (WBAN) has attained a lot of activities in the last years. The reason is the huge activity increase at WBAN applications in healthcare and monitoring. Also standardization procedure has required proper channel models. This paper summary the radio channel measurement and modelling activities for WBAN applications carried out in hospital environment by the Centre for Wireless Communications, University of Oulu, Finland. To fulfil the environmental requirements, the actual measurements were carried out at the Oulu University Hospital premises, in the typical final use places. Different scenarios and link topologies were covered at the measurements. The results show that WBAN channel differs from, e.g., typical office channel models and there is an evident need for aimed channel models for close body communication. I. INTRODUCTION The use of wireless body area networks (WBAN) is a modern way to monitor humans physiological parameters or behavioural changes, and convey the information measured from the body, or even inside a body, seamlessly and imperceptible to remote recipient or electronic database for ubiquitous access. WBAN health monitoring system consists of detectors (sensors) and data communication parts. Depending on the application, a set of sensors used can vary between different patients and their specific health requirements. This causes need to support different kind of traffic and traffic loads inside a WBAN network. On the other hand, dissimilar applications and radio interfaces can allocate different parts of the frequency spectrum at hand. Due to the frequency dependency, and the impact of environment on the propagating signal, accurate and dedicated radio channel models are needed also for WBAN system design. For example, general office channel models do not fit to WBAN propagation environment. Human body is a complex environment due to the close body communications property. A body structure is complex, and human tissues have different electrical properties, both which are impacting on the propagating electromagnetic signal. Moreover, the movement of a body causes extra changes in the specific radio links; for example, in wrist to chest link is varying from line-of-sight link to non-line-of- sight link depending on the phase of the walking cycle and the speed of the person. The WBAN channel realizations and models have an influence from the environment where the wireless body area network operates. There are several existing WBAN channel models available, e.g., [1]-[6]. Also IEEE802.15.6 has defined channel models during its WBAN standardization work [7]. However, there was a lack of WBAN channel models which are based on the measurements carried out at the real hospital environment. Our experimental work has been carried out to fill this gap by providing hospital specific WBAN radio channel models. II. CHANNEL MEASUREMENTS The WBAN channel measurements carried out by CWC covered all together several environments and rooms; anechoic chamber, class room and different hospital rooms. To get realistic environmental features included in the measurement results, three hospital cases were involved in our experiment, namely: operation theatre (surgery room), typical ward room and corridor. The main goal of this experimental work was especially to generate realistic WBAN channel models to be used in designing WBAN applications for hospital usage and to be compared with other created channel models. The measurements were carried out using a Agilent 8720ES vector network analyzer (VNA) [8] having an internal time domain option which made it possible to show a measured response of a radio channel either in a frequency domain (frequency response) or in a time domain (impulse response). The antennas used in the experiments were SkyCross SMT-3TO10M-A antennas [9]. The frequency spectrum covered an ultra wideband (UWB) band between 3.1 GHz and 10 GHz. In the measurements, 100 consecutive frequency responses were measured. The results are based on the statistical analysis of the individual channel responses and their average behaviours. The parameters used during the measurements are shown in Table I. TABLE I MEASUREMENT PARAMETERS PARAMETER VALUE Frequency band 3.1 to 10.6 GHz Bandwidth 6.9 GHz IF bandwidth of the VNA 3.0 kHz Number of points over the band 1601 Maximum detectable delay 231 ns Sweep time 800 ms Average noise floor -120 dBm Transmit power 0 dBm Tx and Rx cables loss 7.96 dB EuCAP 2011 - Convened Papers 3120