Understanding Travel Behavior of Private Cars via Trajectory Big Data Analysis in Urban Environments

Abstract

Private cars, i.e., the vehicles owned for private use, compose a large portion of the civilian automobiles, which play an important role in metropolitan transportation. Private car trajectory offers us an effective way to understand travel behavior of private cars since it is useful in different application areas under urban environment such as path discovery, travel behavior analysis and transportation planning. The existing works regarding trajectory big data analysis mainly concern the floating cars or public vehicles but few consider private cars. In this paper, we focus on studying the travel behavior for private cars based on their trajectory analysis. To achieve this, we investigate the aggregation effect via trajectory clustering with the aim at modeling travel pattern of private cars. We propose a Trajectory Aggregation Detection (TAD) algorithm to find areas where the private cars appear frequently in a fix time interval and then analyze the travel regularity of each individual private car based on trajectory clustering. To validate the proposed method, we have collected large-scale raw dataset of private cars trajectory from real urban environment by installing On-Board Diagnostic (OBD) terminal including motion sensors and GPS receiver. Extensive experiments based on one-year trajectories collected from 1000 private cars reveal that the regularity types of private cars can be identified with high accuracy by the proposed method. We believe that our finding provides a new perspective in studying private car owners' driving pattern and travel behavior.