An Enhanced Indoor Localization System Using Crowdsourced Multi-Source Measurements

Biheng Yang,Bin Li,Lyuxiao Yang,Nan Wu
DOI: https://doi.org/10.1109/iccc49849.2020.9238861
2020-01-01
Abstract:With the rapid development of the mobile Internet, applications based on indoor localization have received increasing attention. In recent years, WiFi received signal strength (RSS) is widely used in indoor localization for the universally available WiFi infrastructure. However, the WiFi signal could easily be affected by non-line-of-sight and multipath propagation, which reduces the localization accuracy. In this paper, we propose an enhanced indoor localization system using multi-source measurements including WiFi RSS, ultra wideband (UWB) ranging, and inertial sensors to improve the performance. The multi-source measurements collected by users' smartphones are used for site survey in our system. To recover users' trajectories, we propose a crowdsourcing method to construct radio map. Moreover, a reference point clustering approach is used to improve system efficiency. A two-step localization method is proposed to locate a user. Experimental results show that the proposed system achieves better performance than only WiFi-based or UWB-based method.
What problem does this paper attempt to address?