Cooperative Vehicle-Actuator System: A sequence-based optimal solution algorithm as tool for evaluating policies

Abstract

Traffic lights are used worldwide for over 100 years. This solution has proved its efficiency especially in urban areas. Nevertheless the complexity of resource sharing problem grows with the increasing number of personal vehicles and the rise of emerging mobility systems. Traffic lights suffer from a lack of flexibility and are not anymore adapted to modern cities. Fortunately the evolution of technology offers efficient positioning and communication systems which permit to imagine a new way to control traffic, the Cooperative Intersection Management (CIM). We have defined a precise protocol named Cooperative Vehicle-Actuator System (CVAS). Real tests have proved the feasibility of the protocol. In this paper we present a sequence-based algorithm based on dynamic programming which aims to find the optimal solution for an isolated intersection. The goal is to have a tool for evaluating real-time policies in simulations. We use a 3D simulator in order to check the efficiency of the algorithm by a comparison with classical traffic lights.