Evaluation of algorithms for UWB indoor tracking

Abstract

Ultra-Wideband (UWB) stands out as one of the most promising technologies for the development of indoor location & tracking systems, providing very precise time of arrival measurement and consequently centimeter-level resolution in distance estimation. Once the distances to multiple reference nodes are estimated, several location & tracking algorithms can be used to compute user's position. Position calculation is a widely studied topic, and many different methods have been proposed in the literature and evaluated under different conditions. Nevertheless, evaluation scenarios are usually generic and too simplistic. The rigorous evaluation of location & tracking algorithms should take into account the specific error distribution of UWB-based distance estimation and the implications of indoor tracking scenarios, such as user's mobility, the number of reference nodes and the distance between them. The objective of this paper is to provide a realistic evaluation of the performance of different location & tracking algorithms (trilateration, least square — multidimensional scaling, extended Kalman filter and particle filter) in indoor environments using a specific UWB-based ranging model to characterize the distribution of distance estimation error.