Exploring Individual Travel Patterns Across Private Car Trajectory Data
Yourong Huang,Zhu Xiao,Dong Wang,Hongbo Jiang,Di Wu
DOI: https://doi.org/10.1109/tits.2019.2948188
IF: 8.5
2020-12-01
IEEE Transactions on Intelligent Transportation Systems
Abstract:Understanding the travel behavior of private cars will generate promising solutions on addressing urban problems such as alleviating traffic congestion and improving transport services. In this paper, we focus on investigating the individual travel patterns of private car users based on a large-scale private car trajectory dataset. To achieve this goal, we first analyze the stop-and-wait information from the private car trajectory data and utilize DBSCAN method to implement clustering with the aim at identifying the frequently-visit places (FVPs). After that, we leverage Markov chain to study the spatial-temporal transition characteristics when private cars travel among their FVPs. Finally yet importantly, we design the concept of spatial-temporal entropy rate and conduct a quantitative study for measuring the regularity of each individual private car's mobility. We validate the proposed methodology based on a real-world dataset including 25,564 private cars driving during one month in China. Extensive experiments demonstrate that the proposed method outperforms the existing methods in terms of the accuracy on measuring the mobility behavior. Moreover, we observe that, on one side, the travel pattern is easier to mine from the private car users with fewer FVPs, on the other side, there are also a small number of users whose FVPs are large, while their mobility are relatively regular. Our work is the first effort to explore individual travel patterns of private car users via studying private car trajectory big data, thereby being able to provide new insight into the research of human travel activities, traffic management and urban planning.
engineering, electrical & electronic,transportation science & technology, civil