Robust TDOA-Based Indoor Localization Using Improved Clock-Sync-Scheme and Multilevel Constrained ARPF
Lixiong Huang, Ruizhi Chen, Ying Chen, Guangyi Guo, Feng Ye, Zuoya Liu, Liang Chen
Abstract
The need for indoor positioning research is driven by the contradiction between the lack of satisfactory location information available indoors and the emergence of an increasing number of indoor location-based services. The ability of ultrawideband (UWB) technology to effectively estimate the signal arrival time and thereby obtain highly accurate ranging information makes it a research hotspot for indoor positioning systems. However, the scalability, user capacity, and ranging performance degradation under nonline-of-sight (NLOS) conditions are challenges that hinder its wide application. To address these issues, this article proposes an improved highly available indoor positioning system that uses a time difference of arrival (TDOA)-based UWB. Within this system, a novel UWB wireless clock synchronization scheme was designed to reduce synchronization errors, improve ranging accuracy, reduce synchronization frequency, and allocate more channel time for positioning. To identify observations with NLOS errors and mitigate their effects on localization, this article implements an adaptive robust particle filtering (ARPF) algorithm based on multiple robustness factors. The multilevel constrained ARPF considers information, such as signal quality, historical observations, and floor map. In a real-world environment, the proposed method is tested using multiple forms of tag devices. The results show that the proposed method outperforms the general particle filter (GPF), cutting the relative error by about the half, especially in the worst badge tests.