Prediction of Traffic Flow at Intersection Based on Self-Adaptive Neural Network

Traffic flow prediction plays an important role in urban traffic management and control. Traditional prediction methods are mostly difficult to meet the high complexity, randomness and uncertainty characteristics of urban traffic flow. In this paper, a new prediction model is proposed based on self-adaptive neural network. Compared with other methods, it possesses the advantages of low computational complexity, fast convergence speed, high goodness-of-fit and so on. Furthermore, it overcomes the drawbacks of vibration effects and easy falling into local minimum caused by single gradient descent algorithms. Simulation results prove the validity of this prediction model.