Ultra-wideband based dynamic target tracking using Cost-Reference Particle Filtering

In this paper, we assess the performance of a sequential Monte Carlo based filter called Cost- Reference Particle Filter (CRPF) in comparison to Extended Kalman Filter (EKF) and Hyperbolic Positioning (HP) based on the time difference of arrival approach for dynamic target tracking. Our results show that CRPF performs better than EKF which in turn performs better than HP. The paper highlights the robustness of CRPF to model inaccuracies which are common in most practical filter implementation problems. The findings of the paper suggest that usage of the CRPF is a promising technique to tackle the problems of random dynamic systems with unknown statistics in ultra-wideband based localization techniques in challenging indoor environments.