Trade-offs in local traffic signal control algorithms on a grid network

Abstract

Turn signal control in intersections play a significant role in intelligent traffic systems. In this work, we study the effect of deploying turn signal controls at varying densities within a synthetic road network. We compare the conventional pre-timed signals to max-queue and max-pressure control algorithms, where timings are adjusted based on the traffic present in the intersection. Increasing the density of controllers encountered in the network significantly increases the travel time of vehicles but can be resolved with the use of the queue-clearing algorithms. At lower controller densities, the use of such algorithms penalizes the minority of vehicles where travel times significantly increase, prominent in the low controller density scenario.