Multipath Components Tracking Adapted to Integrated IR-UWB Receivers for Improved Indoor Navigation
Résumé
In this paper, we propose a solution adapted to Impulse Radio - Ultra Wideband (IR-UWB) integrated receivers devoted to indoor localization, making usage of multipath information under imposed time resolution and low complexity constraints. The Multipath Components (MPCs) can ideally be tracked by capturing their space-time correlation under mobility. The proposed architecture is composed of Multiple Hypothesis Kalman Filters (MHKFs) in parallel, each filter tracking one single MPC while detecting measurement outliers or maneuvering paths. A data association procedure enables to map the MHKFs' outputs onto delays associated with receiver's energy bins. The relative temporal variations of MPCs are used to collectively infer the missing direct path's information in case of Non-Line of Sight (NLoS). The corrected observations finally feed a conventional Extended Kalman Filter (EKF) that estimates mobile's position. We evaluate the performance of the proposed scheme through realistic simulations in terms of both MPC and mobile tracking.