Evaluation of Flow-Based Traffic Signal Control Using Advanced Detection Concepts

Abstract

State-of-the-practice signal timing techniques used at isolated actuated controlled intersections can result in suboptimal operation. This paper proposes a new tactical control algorithm that integrates real-time stop bar presence detection with real-time flow rate information to identify a downstream flow restriction. If a flow restriction is identified, a phase with a constant call could be terminated earlier than its specified maximum or split time. This algorithm was validated with real-time data to predict downstream bottleneck conditions and then by viewing the archived video for confirmation of the condition. The algorithm was applied during a 504-h period of signal operation. During this period, the algorithm identified 76 bottlenecks. Overall, 75.0% of these events were visually confirmed. Algorithm failures were caused by a variety of conditions such as large trucks, inattentive drivers, and detector errors.