Intelligent traffic control based on a combined model of Fuzzy logic and Reinforcement learning

Increasing the number of vehicles on the streets is called the problem of urban traffic congestion. One way to solve this problem is to control the timing of traffic lights. In this research, the model used is the green-red space model and the yellow light as a third color has been added to the modeling. To control the illuminated intersection, a fuzzy amplifier-learning controller is used, the core of which is the Fuzzy Q-Iteration algorithm. The length of each street queue is considered as a fuzzy variable. The controller generates a control signal according to the length of the queue behind the light. The output control signal is the duration of the green light on each street during a cycle. The results show that the proposed controller had a similar or better performance than the fixed time controller ratio with the vehicle waiting time criterion. At high input current rates, controller performance has improved significantly in reducing waiting times. In addition, the queue length on streets with high input flow rates is reduced because the agent tries to generate a larger control signal on high flow rates streets, which means more green time for that street. According to the proposed model, the number of cars on each street of the smart intersection does not exceed about 30 cars.