Format

Send to

Choose Destination
Sensors (Basel). 2018 May 31;18(6). pii: E1758. doi: 10.3390/s18061758.

An Enhanced Hidden Markov Map Matching Model for Floating Car Data.

Author information

1
School of Geographic Science, Nantong University, Nantong 226019, Jiangsu, China. dawnche@163.com.
2
School of Geographic Science, Nantong University, Nantong 226019, Jiangsu, China. wyl621021@ntu.edu.cn.
3
School of Geographic Science, Nantong University, Nantong 226019, Jiangsu, China. benz1983@163.com.
4
School of Geographic Science, Nantong University, Nantong 226019, Jiangsu, China. cxliang@mail.ustc.edu.cn.

Abstract

The map matching (MM) model plays an important role in revising the locations of floating car data (FCD) on a digital map. However, most existing MM models have multiple shortcomings, such as a low matching accuracy for complex roads, long running times, an inability to take full advantage of historical FCD information, and challenges in maintaining the topological adjacency and obeying traffic rules. To address these issues, an enhanced hidden Markov map matching (EHMM) model is proposed by adopting explicit topological expressions, using historical FCD information and introducing traffic rules. The EHMM model was validated against areal ground dataset at various sampling intervals and compared with the spatial and temporal matching model and the ordinary hidden Markov matching model. The empirical results reveal that the matching accuracy of the EHMM model is significantly higher than that of the reference models regarding real FCD trajectories at medium and high sampling rates. The running time of the EHMM model was notably shorter than those of the reference models. The matching results of the EHMM model retained topological adjacency and complied with traffic regulations better than the reference models.

KEYWORDS:

floating car data; hidden Markov model; map matching model; satellite positioning systems; topological adjacency; traffic regulation

Supplemental Content

Full text links

Icon for Multidisciplinary Digital Publishing Institute (MDPI) Icon for PubMed Central
Loading ...
Support Center