Study And Development Of Speech Control System In Assistive Bath Appliances For Persons With Disabilities

With the social progress, the life quality of human beings improved much. In themeanwhile, people become more and more busy. As a population aging society like China,we could not throw in too much manpower to take care of the old and disabled. So,assistive appliances for disabled people can create enormous social benefits and have ahuge potential for development in the future. Besides, intelligence is a future trend forproduct, and pattern recognition is an important part of intelligence. In our study, weintroduce the speech recognition technology to the assistive appliances, and bring up aviable solution.Based on our study, we developed a speech control system for the assistive bathappliance for persons with disabilities. Firstly, we made some needs analysis of theassistive bath appliance, then we studied the speech signal processing method based onsignal modeling theory. And we compared the existing characteristic parameters andpattern recognition technologies, demonstrated the advantages and disadvantages of them,then we brought up a scheme using Mel-frequency cepstrum coefficient as characteristicparameter and dynamic time warping as recognition method. With respect to the specificnoise in the bathroom, we introduce a modified spectral subtraction algorithm, use it insome experiments and analyze the result.Based on the former Scheme, we have developed a multi-thread speech recognitionsoftware in Windows platform by Visual C++6.0integrated development environment. Ituse muti-thread to realize parallel processing of speech signal sampling and processing. Atlast, we test the software and show some results in the conclusion.