An advanced UWB channel model for body-centric wireless networks

This paper presents a novel ultra wideband (UWB) channel model in the 3-10 GHz range for body-centric wireless communications. The tests are performed in both indoor anechoic chamber environments, addressing on-body and off-body propagation scenarios. The body channel model is extracted by using a single spatial grid over all the body, and by distinguishing between LOS and NLOS condition. The large number and the uniform placement of the receiver locations attempt a representation of the body propagation links more comprehensive than previously published models. The statistical reliability of the model is investigated by applying jointly the Kolmogorov-Smirnov and the Akaike criteria. The analysis suggested that the Lognormal model fits the channel amplitude distributions with a percentage ≥ 64%. The on-body indoor channel amplitudes are modeled with a stochastic terms of about 4-5 dB higher than previously published models. Finally, a Negative-Binomial and Inverse Gaussian distribution are used to model the expected number of paths and interarrival time, respectively. Based on the results presented in this paper, clear recommendations are given with regards to the optimum statistical distribution of an accurate UWB body-centric radio channel modeling.