Understanding and Managing Travel from a New Perspective: A Study on Vehicle Usage Characteristics Based on One Week of Spatiotemporal Trajectory Data from Shanghai

Abstract

Exploring the usage characteristics of private vehicles is the basis of traffic demand forecasting and management. However, it has been difficult in previous research due to lack of travel data for continuous days. The boom of electric vehicles (EVs) makes vehicle travel behavior more complicated, and on the other hand, the trajectory data for multi-days generated from EVs brings opportunities for a deeper understanding of vehicle activity features. In this paper, based on a week’s operation data of 8,000 household EVs in Shanghai, we design an algorithm to identify the stop points and their types (residential, working, and other places). Based on the relationship between vehicle travel characteristics and the types of vehicle stop points, we classify vehicles into four categories using a GMM model: the family-used commuting (CMT) vehicles, the family-used non-commuting (Non-CMT) vehicles, the commercially used (CMR) vehicles and the semi-commercially used (Semi-CMR) vehicles. A new method is proposed to identify the regular trips and we use the Z Test to discuss travel time and route choice differences between four kinds of vehicles. We find that CMR vehicles present high travel intensity in temporal and spatial aspects and the use intensity in elevated roads is higher for CMT vehicles than Semi-CMR vehicles. The methodologies and conclusions of this paper not only provide theoretical support for future urban traffic prediction but guidance for employing refined active traffic demand management measures to alleviate traffic congestion.

Keywords: electric vehicles, location identification, vehicle classification, travel behavior analysis, traffic demand management