Artificial neural network estimation of vehicle audio signal in high way

In this paper, during traffic flow information signal extraction and analysis of the sound of cars on the freeway, the linear transfer function of three-layer artificial neural network was used for the hidden layer neurons. Appropriate AR model order selection can distinguish different models in order to find the feature vector. Experimental results show that the estimation AR model parameters and power spectrum of freeway vehicle audio signal using the neural network method is effective.