An Adaptive Wavelet Neural Network for Traffic-Flow Prediction

Traffic-flow prediction is an important issue for the traffic management work, especially in the big cities. Although many prediction models are studied and used to predict the traffic-flow in several application scenarios, they do not perform well in the computational complexity and robustness. In this paper, we propose an adaptive wavelet neural network (AWNN) for traffic-flow prediction. In the proposed AWNN model, a wavelet neural network (WNN) is firstly constructed and driven by the traffic-flow data. Moreover, an adaptive learning algorithm with two-step ahead information is designed in the process of updating parameters to further improve the adaptive ability. Finally, the simulation experiments on the real traffic-flow data are conducted to verify the effectiveness of the proposed AWNN.