On-body TOA-based ranging error model for motion capture applications within wearable UWB networks

We present herein an error model that characterizes on-body range measurements based on time of arrival (TOA) estimation in Impulse radio-ultra wideband, wireless body area networks. Considering real channel measurements over two representative on-body links for repeated walk cycles, the model is drawn as a conditional mixture, accounting for signal to noise ratio (SNR) variations and non line of sight (NLOS) channel obstructions caused by the body. Key model parameters are then investigated as a function of the previous obstruction and SNR configurations, illustrating missed/false path detection effects at low SNR. On this occasion, two TOA estimators are compared, namely a strongest path detection scheme through matched filtering and a first path detection scheme relying on high-resolution channel estimation. Finally, we discuss the possibility to generalize the previous model to any kind of on-body link, based on empirical observations regarding the dynamic range of the channel power transfer function under mobility. Accordingly, the resulting final model could integrate basic elements of classification, such as the instantaneous LOS/NLOS and static/dynamic link status.