Hierarchical Fuzzy Logic Control for Multiphase Traffic Intersection Using Evolutionary Algorithms

In this paper, a hierarchical fuzzy logic traffic controller is constructed for a real intersection of fourteen vehicle lanes and two pedestrian crossings controlled by signals with seven light phases. The hierarchical fuzzy controller has seven inputs as queue lengths of the seven light phases, and one output as green time of the selected phase. In the hierarchical fuzzy controller, there are six layers of fuzzy sub-controllers with two inputs and one output. The sub-controllers in the first five layers have identical structure that has two inputs of queue lengths and one output of combined queue length employed as one input of next layer. The sub-controller in the last layer has two inputs, combined queue length obtained from the fifth layer and queue length of the selected phase, and one output as green time of the selected phase. Using the developed fuzzy controller, the best fuzzy rule base is obtained based on real traffic data of the intersection by employing evolutionary algorithm. performance of this controller is simulated and compared to that of a controller that is currently employing in the intersection. The results show that the developed fuzzy controller shortens more than 38% of the vehicle waiting time.