Edge computing can reduce delay of computing, data transmission bandwidth and alleviate pressure of cloud computing, which brings many opportunities for upgrading existing intelligent transportation systems. This paper proposes an edge computing system to efficiently track hazmat transport vehicles. In the system, multi-source data relevant to hazmat transport vehicles are collected, including data from on-board recorders, video from smart cameras and electronic license plate information from electronic vehicle identification devices at roadside. During the tracking, GPS trajectories are used for primary tracking source, while other data are processed at the edge to surveil the targeted vehicles and evaluate their status. Alarm can be raised automatically once anomaly of hazmat transport vehicles is detected. An efficient cooperative mechanism among edge nodes is devised to reduce the network bandwidth and computational requirement. For evaluation, a prototype tracking system upon SUMO is implemented. The experimental results show the efficiency of our system compared with conventional cloud-based systems.

CICTP