Application of Speech Recognition Using Cepstral Weighting and Time Normalization In Multi Output On-Off Controller

This paper addressed the multi output controller that activated by speech command design and realization. Linear Predictive Coding is used as the method of speech signal processing and Dynamic Time Warping is used to recognise speech pattern and to normalize time. The controller can has outputs tip to 256 outputs. Each output can be used to drive an external device or can he combined as logic controller. Experiments show that percentage of the average success recognition is 98.16 %. 7hi.s result obtained by using speaks speed variation. F'or normal and slow speak the percentage of .success recognition is 100 %. And percentage of rejection of the words, which not in the reference pattern is 63.3% Key words: speech, recognition, multi output controller, DTW