Deep Q-Network Based UWB Anchor and Strategy Selection for Accurate Pedestrian Localization in Vehicular Environments

Ultra-wideband (UWB) technology is emerging as a key technology for indoor navigation and positioning systems, offering reliable, accurate, and energy-efficient performance. This characteristics of UWB can be leveraged to localize pedestrians in vehicles equipped with UWB technology. However, the localization of pedestrians in a vehicle environment, which is a more complex scenario than an indoor environment, requires further consideration. In particular, the selection of a line-of-sight (LoS) link and the delusion of pecision (DoP) problem are of paramount importance for the performance of the localization process. To address these issues, we propose a deep Q-network based UWB anchor and localization strategy selection that improves the localization success rate by identifying LoS links without relying on complex LoS/NLoS identification methods. Additionally, it determines the optimal localization method and addresses DoP issues to enhance location accuracy. The proposed scheme is evaluated against existing techniques in terms of localization precision at varying distances, LoS/NLoS conditions, and scenarios involving vehicle and pedestrian moving. The results demonstrate that the proposed approach outperforms existing methods.