Fuzzy Graph-Based Controller for a Real-Time Urban Traffic Optimization

Abstract

This paper discusses the development of an intelligent system to remedy the problem of urban congestion. Thus, both fuzzy logic and graph theory are adopted to achieve this aim. The fuzzy logic controller generates the weights between two vertices of a graph, which constitutes a basic modeling of an urban road traffic. A multi-criteria analysis allows synthesizing simple and efficient fuzzy rules based on human expertise. These criteria can be divided into two basic specifications, which are the number of cars present on a road, and the distance between two intersections that limit this road. Furthermore, three additional criteria are used, namely, the detection of accidents, the presence of public works and the maximum speed permitted. The elaborated system can be used to model any urban traffic network with a simple adjustment of membership functions and the discourse domain of the fuzzy system. The weights generated by the developed fuzzy system are between zero and one. The weight converging towards zero can be considered the optimal one. The simulation using MATLAB/SIMULINK shows the efficiency of the method developed for fixed and dynamic inputs, and it shows the effectiveness of the fuzzy rule base in managing changes in entry criteria.