Urban Traffic Signal Learning Control Using SARSA Algorithm Based on Adaptive RBF Network

Urban traffic control is very complicated, so to build a precise mathematical model for it is very difficult, In this paper, we use the SARSA reinforcement leaning algorithm to control the traffic signal, thus the decision can be made dynamically according to real-time traffic state information, and the change of environment can be adapted automatically; As the state space is too big to be stored and expressed directly, we applied radial basis function neural network (RBF) to approximate the state value function. By training self-adapted non-linear processing unit, and realizing online and adaptive constructing of state space, the approximation is improved and thus the control of traffic signal at single crossroad is solved. The simulation results show that the effectiveness of the new control algorithm is obviously better than traditional fixed time slot allocation method.