On the sparse non-parametric model for body-centric ultra-wideband channel

Parametric probability models are usually used to characterize the channels for Wireless Body Area Networks (WBANs). However, one should note that only limited samples can be collected in WBANs; on the other hand, channels for this area can be easily affected by many unexpected factors, such as human body postures, individual difference, etc. In this paper, the authors develop a sparse non-parametric model for UWB WBANs, it is demonstrated that when the coefficients are selected properly, the error of the model is smaller than traditional parametric ones.