Applying agent based simulation to the design of traffic control systems with respect to real-world urban complexity

Abstract

The problem of reducing traffic congestion in a city has always been difficult to solve with monolithic control methods, which have both high costs and increased implementation complexity. This paper aims to minimize vehicle waiting time at stoplights by using a multi-agent system control technology. Moreover, the system is required to respond adequately to the presence of emergency intervention vehicles, allowing them quick and sure passage, but without significantly interrupting regular traffic. The solution designed in this paper allows for on demand synchronization of intersections, depending on the traffic context at any given time. In order to test this concept, an agent based simulation model has been developed, that offers real world traffic simulations on urban maps, and integrated complex road networks and traffic participant behaviour, with a possibility to measure the performance of the control system through parameters such as noise levels and emissions.