A Carpooling Service for Private Vehicles Using Electronic Registration Identification Data

摘要

Carpooling of private vehicles can obviously reduce urban road traffic volume and traffic jams. This study aims to mine travel behavior of private vehicles and establish a stable carpooling relationship between private vehicles. We use electronic registration identification (ERI) data to discuss private vehicles' travel. ERI is a means of uniquely identifying vehicles, which can cover all vehicles including private vehicles and better reflect the city's traffic overall characteristics. In this paper, we propose a method based on data cubes to mine regular travel behaviors of private vehicles for carpooling services. Then, the degree of recommendation between owners of private vehicles is obtained by measuring their spatiotemporal similarity of trip, similarity of vehicle and similarity of origin. It is believed that users with a higher recommendation degree have the possibility of carpooling. Finally, we use the real-world data, which is generated by 1,090,136 private vehicles in a working week in the city of Chongqing, China, to do experiment. Results show that carpooling services can reduce 6.09% private vehicles and reduce 3.35% mileage during morning commuting period.