Body Area Communications: Channel characterization and ultra-wideband system-level approach for low power

Demographic trends are leading to increasing health care costs and fewer hospital beds creating demand for technologies that move health care out of the hospital while minimizing pressure on the family and community. Wireless sensor networks worn on the body to remotely monitor patients and administer drugs have recently been proposed to meet this demand. At the same time, regulatory agencies have released an unprecedented amount of bandwidth for ultra-wideband (UWB) communication to meet the requirements of future wireless services. Despite these encouraging developments, the full potential of UWB wireless body area networks (WBANs) for personal health monitoring has never been realized. This is because the body area radio channel is poorly understood making it impossible to evaluate WBAN communication systems. To address this problem, we have first derived an analytical body area propagation model directly from Maxwell's equations. By considering the body as an infinite lossy cylinder and the antenna as a point source, we have predicted the basic propagation trends expected around a human body and confirmed this with anechoic chamber measurements. Based on these insights, we have conducted an extensive measurement campaign to develop physically-motivated statistical propagation models for both narrowband and ultra-wideband systems. These models not only capture the influence of the human body, but also the contribution of nearby scattering objects. The narrowband model uses simple pathloss laws and fading distributions common in traditional wireless systems, firmly establishing WBANs within the existing propagation modeling and analysis framework. Similarly, we have shown how tapped delay line models can be used to study ultra-wideband body area propagation. Specifically, we have proposed a novel modeling form where components diffracting near the body are measured and modeled separately from later arriving reflections. This modeling form is more appropriate for capturing the different stochastic behavior of these separate contributions. Based on our body area model, we have shown how a mostly-digital sub-sampling UWB radio architecture that minimizes the number of analog components and effectively exploits technology scaling offers a promising solution for low cost WBANs. Sub-sampling architectures were never widely deployed in conventional systems due to noise aliasing and sampling jitter. However, the wide bandwidth of UWB makes it less susceptible to noise aliasing. Furthermore, we have derived the precise relationship between sampling jitter and communication performance demonstrating that UWB is also robust to practical jitter tolerances. Finally, we show how even low-complexity digital baseband solutions for acquisition, Hilbert transforms, pulse shape estimation, and clock offset tracking result in less than 2dB implementation loss in worst-case body area propagation conditions. Based on this performance analysis and the power consumption of recent UWB prototypes, this dissertation concludes by showing how emerging UWB technologies can dramatically improve both the range and battery lifetime of WBAN systems compared to existing standardized narrowband solutions.