Joint motion capture and navigation in heterogeneous Body Area Networks with distance estimation over neighborhood graph

We present and compare here solutions adapted to heterogeneous Wireless Body Area Networks (WBANs) to jointly fulfil indoor navigation and motion capture needs over large-scale trajectories. On-body devices are localized at the building scale by combining peer-to-peer range measurements over intra-WBAN links and further measurements with respect to external anchors forming the building infrastructure. Those measurements rely on two main low data rate radio technologies, namely the Impulse Radio - Ultra Wideband (IR-UWB) and a Narrow Band technology at 2.4GHz, providing respectively Time (Difference) Of Arrival (T(D)OA) estimates and Received Signal Strength Indicators (RSSI). This new Large Scale Individual Motion Capture (LSIMC) functionality is then investigated within two algorithmic embodiments. In a first 2-step scheme, the positions of on-body nodes are preliminary estimated in a body-strapped Local Coordinates System (LCS) before being expressed into an absolute Global Coordinates System (GCS) after a set of transformations, whereas the second option considers localizing directly the nodes into the GCS using all the available measurements. Simulations illustrate particularly the harmful impact of body obstructions. We also propose a method to mitigate these effects and approximate measurements over obstructed or erroneous links, by computing the shortest distance in graph neighborhood.