Automated Traffic Signal Performance Measures (ATSPMs) in the Loop Simulation: A Digital Twin Approach

Abstract

In this paper, we present a digital twin approach to the enhancement of traffic signal performance monitoring and congestion identification using automated traffic signal performance measure (ATSPM) systems. The objective of this effort is to use the high-fidelity microscopic simulation engine to generate simulated traffic signal events and connected vehicle data and allow the ATSPM systems to generate various traffic signal measures of effectiveness (MOEs) for a forensic traffic signal evaluation. Real-world ATSPM systems are driven by the traffic signal logs generated during operations. Therefore, they are primarily applied to traffic signal operations in the field. However, traffic signal design at present still follows the traditional method based on averaged delays, stops, and so forth, while more and more agencies have begun to evaluate the implemented traffic signal systems’ performance using the novel ATSPM MOEs. The proposed ATSPMs-in-the-loop simulation system fills this gap by using the ATSPM systems to evaluate the proposed traffic signal timings at the design stage. The benefits of this system include providing full-spectrum decision support for traffic signal management from design to operation and facilitating agencies to develop new insights on identifying traffic signal problems using the ATSPM MOEs. Another feature of the ATSPMs-in-the-loop simulation system is that it can use the emerging connected vehicle data set to generate new traffic signal MOEs. In the case study, we demonstrate how to use the proposed system to identify the potential issues of detector layouts and bottlenecks. Additional features of this ATSPM digital twin include allowing external components to interact with this platform via standard protocols in traffic control systems and connected vehicles to serve more purposes.