Adaptive traffic lights based on hybrid of neural network and genetic algorithm for reduced traffic congestion

Traffic congestion is a challenging problem in the present scenario where we are enjoying the conveniences of automobiles every day and want faster transportation. This problem is increasing exponentially day by day so to deal with this problem we devise an adaptive traffic signal controller (TSC) as traditional traffic signal controllers are inefficient in dealing with increasing demands of growing traffic. This controller uses neural network (NN) and Genetic Algorithm (GA) to adapt the traffic signal timings according to the congestion. NN takes signal timings as input and gives the queue length as output. GA is further applied to get the optimized green signal timing at its output, which is capable of reducing the queue length and overall delay. The performance of proposed model is also compared with fixed time TSC and an already existing adaptive TSC and a significant improvement were observed.