Signal Coordination with Simple Agent-Based Rules

Abstract

In todays road networks, traffic management methods are of more and more importance with steadliy increasing traffic. Detector values are related to preassigned, fixed signal programs or used to adapt the signal phases within a signal phase frame. Drebacks of these detectors are high installation costs and limitations in data usage. From a single detector information on vehicle speeds or number of vehicles waiting is not available. With the development of information technologies new kinds of detectors aloow a more flexible operation of traffic signals by using additional information on traffic flow. This allows in turn the development of concepts of autonomous actin traffic lights reacting on the demand at the intersection or within the approaches. We give an example of subsequent autonomous acting traffic lights equipped with a simple rule set optimizing the traffic flow along a road segment. The rule set is based on a cross correlation between detected vehicles in a certain distance from the traffic lights and the green time. The optimal phase shift with respect to parameters of flow, mean vehicle speed and traffic light position is discussed.