Digital Twin of Physical Intersection to Trajectorybased Traffic Signal Controller

Commercially available signal controllers mostly use data from loop detectors in the form of binary contact closure. With the emergence of technology, advanced sensors, e.g., RADAR, LiDAR, or video cameras, can track vehicle position and speed data continuously. However, none of the commercial signal controllers can explicitly integrate vehicle trajectory data into the core control logic. We have developed a signal controller tentatively named “Green Box Signal Controller (GBSC)” to integrate both loop detector and trajectory data from sensors explicitly. Trajectory data from advanced sensors are primarily presented in the form of cartesian coordinates, and this requires further processing to create a digital twin of the actual intersection to the trajectory-enabled signal controller. In GBSC, we have used an approach-lane pair structure to create a digital map of physical intersections. In this paper, we explained how GBSC works and how intersection layouts can be mapped in GBSC. Furthermore we have validated performance of GBSC with Econolite ASC/3 controller.