Study of substation inspection robot voice recognition algorithm based on wavelet transform

Abstract

A kind of substation inspection robot device voice recognition algorithm is proposed based on wavelet decompositions, which can recognize the running state of substation device. Firstly, the sample library was collected by robot pickup. Secondly, the sub-band energy of samples was extracted by using wavelet decomposition; the sample code book was billed by using VQ algorithm and LBG algorithm. Finally, using the codebook to identify sound samples, which can judge the running state of the equipment. The experimental results show that, the algorithm can effectively identify the sound equipment and high accuracy. The requirement of the substation inspection robot for detecting device is satisfied.