Traffic information extraction of vehicle acoustic signal based on wavelet packet analysis

A method for traffic information extraction of vehicle acoustic signal based on wavelet packet analysis is proposed. At first, the basic principle of wavelet packet analysis is introduced, and uses db5 wavelet function to decompose the acoustic signal by three-layer wavelet packet. According to the characteristic of the vehicle acoustic signal frequency which is in low frequency band, the sub-band energy at node (3, 0) is considered as the key parameter of the signal. At last, a threshold of classification is proposed based on the analysis of the experimental data, and the validity of this method is verified through large number of experiment. The recognition accuracy could be averagely reached 92%.