UWB Channel Modeling for BANs based on the ARMA Synthesis Method

Abstract—As UWB channels for BANs behave as linear systems, an alternative form of channel modeling is by linear prediction of the channel’s transfer function obtained from measurements. The main reason of this is a more accurate reproduction of the channel transfer function with less parameters than conventional statistical channel models. In contrast to conventional statistical channel models, where channel impulse responses are constructed from a statistical characterization of multipath amplitudes, phases and delays arriving in clusters, etc., the channel impulse response derived from an ARMA channel model is more accurate than a statistical channel model. Indeed, as this ARMA channel model is derived directly from measured channel transfer functions. So, it is expected that this channel impulse response is more accurate. Recall, a statistical channel model has some degree of error depending how precise the statistical characterization was performed. Moreover, the whole propagation phenomena (large scale and small scale fading) are embedded in a simple model. Index Terms—UWB channel modeling, body area networks (BANs), auto-regressive moving-average (ARMA) synthesis method. I