Map-matching approach based on link factor and hidden Markov model

Article type: Research Article

Authors: Zhang, Haiyan^{a; b} | Luo, Yonglong^{a; b; *} | Yu, Qingying^{a; b} | Zheng, Xiaoyao^{a; b} | Li, Xuejing^{a; b}

Affiliations: [a] School of Computer and Information, Anhui Normal University, Wuhu, Anhui, China | [b] Anhui Provincial Key Laboratory of Network and Information Security, Wuhu, Anhui, China

Correspondence: [*] Corresponding author. Yonglong Luo, E-mail: ylluo@ustc.edu.cn.

Abstract: An accurate map matching is an essential but difficult step in mapping raw float car trajectories onto a digital road network. This task is challenging because of the unavoidable positioning errors of GPS devices and the complexity of the road network structure. Aiming to address these problems, in this study, we focus on three improvements over the existing hidden Markov model: (i) The direction feature between the current and historical points is used for calculating the observation probability; (ii) With regard to the reachable cost between the current road section and the destination, we overcome the shortcoming of feature rarefaction when calculating the transition probability with low sampling rates; (iii) The directional similarity shows a good performance in complex intersection environments. The experimental results verify that the proposed algorithm can reduce the error rate in intersection matching and is suitable for GPS devices with low sampling rates.

Keywords: Map-matching, hidden Markov model, link factor, reachable cost, directional similarity

DOI: 10.3233/JIFS-202292

Journal: Journal of Intelligent & Fuzzy Systems, vol. 40, no. 3, pp. 5455-5471, 2021