ITS165222智能交通世界大会ITS智慧城市社区人工智能AI物联网IT报告课件教案.pdfVIP

22nd ITS World Congress, Bordeaux, France, 5–9 October 2015 Paper number ITS-1652 A Robust WiFi Fingerprinting-Based Vehicle Positioning System For Heterogeneous Devices 1* 1 1 1 1 Wang Lei , Umer Rasheed , Ng Hai Heng , Ang Chee Wei , Xu Wenjie 1. Intelligent Transport Systems, Institute for Infocomm Research, A*STAR, Singapore. 1 Fusionopolis Way, #21-01 Connexis, wanglei@i2r.a-star.edu.sg Abstract Many Received Signal Strength (RSS)-based fingerprinting positioning techniques have been recently proposed to exploit the increased coverage and availability of WiFi access points (AP) in both indoor and outdoor environments. One practical challenge in the fingerprinting approach is on the device diversity problem and several prior works have attempted to mitigate this adverse effect. However, these existing algorithms can only work well for indoor environments. In this paper, we propose a novel power-gap elimination (PGE) algorithm for outdoor vehicle positioning. Our experimental results show that the fingerprinting-based vehicle positioning is also feasible even in the presence of device diversity, which can be handled by incorporating the PGE. Moreover, we also propose a new method to obtain the level of accuracy for a location estimate in a real-time manner, so that a location estimate can be more informative. Keywords: vehicle positioning, fingerprinting, heterogeneous devices. I. Introduction In the recent past, numerous applications have emerged that rely on accurate and continuous location information [1]. Global Positioning System (GPS) is the primary source of l