Development of multi-agent ANFIS-based model for urban traffic signal control

In this paper, an adaptive neuro-fuzzy inference system (ANFIS) based model for urban traffic signal control was developed. The ANFIS-based control scheme adaptively controls phase sequence and phase lengths to ensure smooth flow of traffic, decrease traffic delays and hence enhance effective road usage. In this design, a multiagent approach was adopted, and an agent (ANFIS-based) used traffic information such as queuelength and waiting time in an intersection, and queuelengths information received from agents of neighboring intersections, to effectively and efficiently control traffic in real-time at any given intersection. A simulator is developed using MATLAB/SIMULINK software for a network of eight intersections in Uyo Metropolis of eastern Nigeria as a case study. Performance evaluation results showed that for three different traffic volume scenarios considered, ANFIS-based traffic control scheme significantly outperforms the existing and optimized fixed-time controls in terms of delay, throughputs and queuelength.