Speech recognition algorithm of substation inspection robot based on improved DTW

Abstract

The voice recognition algorithm is proposed in this paper which can control the substation inspection robot by voice, so the voice-controlled robot is realized. Combined the voice controlled command collected by robot’s pickup device and the basic principle of speech recognition, the intelligent mobile robot can make the corresponding action according to the instructions and complete automatic detection and information query. Firstly, the voice instruction was collected by using robot and transmitted system based to construct the sample database. Secondly, the sample database was analyzed and MFCC feature of speech sample was extracted. Finally, the template matching of voice parameters was achieved by using the improved DTW as the matching algorithm, and the recognition result was transmitted the robot system to control the robot’s action. The experimental results show that the algorithm can quickly recognize and extract the voice command, so it can improve the recognition accuracy and real-time control the inspection robot.