Control of a Wireless Robot Using Speech Recognition in Punjabi Language

Abstract: Speech recognition and analysis have made great progress, over the last few years. Speech recognition technology make possible creating voice controlled robots. In this paper, a speech recognition system is developed by using Mel Frequency Cepstrum Coefficient (MFCC), Vector Quantization (VQ) and Artificial Neural Network (ANN) to control of a wireless robot by using regional language (Punjabi). The commands that used here are all in Punjabi language. The voice signal for both male and female are recorded in .Wav file at 16 kHz sampling rate and then modified. The features of collected voice signal are extracted by using MFCC and VQ‟s. Back propagation method is used to train artificial neural network. The training data include the pronunciation of six words used as the command. These commands are created from 25 people including both male and female. 750 data training is used to train artificial neural network.

Keywords: Robot, Voice Commands, Speech Recognition, MFCC, VQ‟s and Artificial Neural Network.