Vehicle Localization and Velocity Estimation Based on Mobile Phone Sensing

• Published in: IEEE access Vol. 4; pp. 803 - 817
• Main Authors: Chen, Chi-Hua, Lee, Chi-Ao, Lo, Chi-Chun
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2016; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Cell phones; Cellular phones; Cellular telephones; Electronic devices; Global Positioning System; Global positioning systems; GPS; IEEE 802.11 Standard; Intelligent sensors; Localization; Log normal distribution; Mobile communication; Mobile phone sensing; Normal distribution; Position measurement; positioning algorithm; Power consumption; Power distribution planning; Power management; Power supplies; received signal strength indicator; Sensor systems; Sensors; Signal strength; Sustainability; Wi-Fi-based localization; Mobile phone sensing; positioning algorithm; Wi-Fi-based localization; log-normal distribution; received signal strength indicator
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2016.2530806

In recent years, smart devices have become ubiquitous. Many of these devices are equipped with global positioning system (GPS), Wi-Fi, and other sensors. With their high mobility, the idea of mobile devices as probes has been attracting more and more attention. Mobility and flexibility offered by smart mobile devices are what traditional fixed sensors lack. However, mobile devices' power supplies are quite limited. Although GPS is accurate, its high power consumption somewhat limits its accessibility and sustainability. In contrast, Wi-Fi is less power hungry, but at the same time, less accurate. For the sake of sustainability, by adopting Wi-Fi as an alternative to GPS, longer operation is attainable at the cost of losing some degree of accuracy. In this paper, a Wi-Fi-based algorithm based on log-normal probability distribution of distances with respect to received signal strength is proposed. It is suitable for an outdoor environment, where Wi-Fi access points (APs) are abundant. Simulations are conducted over known AP locations, and results show that the proposed algorithm can save, on average, as much as 35% more battery power than GPS does. The average localization error is ~18 m, and the average velocity estimation error is ~25%.