Red light running prediction for dynamic all-red extension at signalized intersection

This paper reports on a red-light-running (RLR) prediction scheme for an intersection collision avoidance system that dynamically extends the all-red phase to avoid RLR related collisions. The paper addresses the issues of predicting the red-light running using an infrastructure-based sensor and a vehicle-to-roadside communication (Intellidrive) system. In the infrastructure-based sensor case, we validate the performance of the improved prediction algorithm that incorporates the inter-vehicle (car-following) status. From a field experiment, the study shows that the accurate and timely RLR prediction potential can be obtained through the Intellidrive-enabled system. Moreover, the perspective of the integration of infrastructure-based sensor and Intellidrive is also analyzed in this paper.