Evaluation of the impact of traffic incidents using GPS data

Abstract

This paper proposes a trajectory of vehicles encountering an incident site and offers six indices for evaluating the impact of the incident on traffic patterns based on global positioning system (GPS) data. Traffic incidents are the major contributors to non-recurrent delays. Mitigating their effects necessitates an appropriate and efficient allocation of equipment and personnel for a clearance plan. Reliable real-time traffic impact evaluations are essential prerequisites for this task. Most of the existing duration models, however, are developed on the basis of local data, and may not be portable because of inconsistent data structures and differing local conditions. The method proposed in this paper is not limited by these shortcomings because it is based purely on classical traffic flow theory. The proposed trajectory is calibrated and validated using real-world taxi GPS data and accident data in Hong Kong. The proposed indices of severity are estimated through analysing the calibrated trajectories. The results reveal that the proposed indices can reconstruct the scene of an incident site and that the proposed method is promising. With the widespread use of GPS technology, the potential benefits of this method for incident management can be realised.