Researches And Applications Of Speech Recognition Technology In The Smart Home

With the progress of human society and the rapid development of science, people began to pursue the smart and convenient home environment, and therefore smart home is quickly rising. Smart home is based on family residential, and it integrates many facilities of home life via computer communications, pattern recognition, automatic control and other technology to build an efficient, smart home automation control and management system. In traditional smart home system, control modes include manual control and remote control. And with the rapid development of science and more needs of user experience, speech recognition and voice control technology is more and more important in the smart home. Speech recognition technology recognizes the semantic information of speech, language and speaker information from speech signal. Speech recognition technology is one of the important areas in the speech signal processing. It is also an important branch of pattern recognition.Currently it has important theoretical value and broad application prospects.The background of this paper is based on smart home, and the speech recognition technology is applied to the smart home environment. Commands phrases are recognized by speech recognition technology to control home appliances. This paper has built an efficient, smart home automation control and management system and made the system more intelligent. This paper is focused on the specific people’s phrase speech recognition and how to judge the control commands phrases.To be more specific, the main works are as follow:(1) In the aspect of speech feature analysis, this paper analyses time domain, frequency domain and cepstrum domain of the speech phrase signal. Speech phrase signal’s feature is analyzed by some experiments. This paper do speech signal processing experiments and get the analysis results by the wavelet analysis technology.(2) In the aspect of speech recognition, dynamic time warping algorithm is applied for recognition algorithm about commands phrases. The recognition results are analyzed from different experiments’views.(3) In the aspect of recognizing the voice control commands, Semi-supervised learning is applied for classifying and recognizing to process the speech recognition results. This paper uses Transductive support vector machines algorithm and Transductive support vector machines algorithm with Deterministic Annealing technology, semantic recognition accuracy has improved.(4) In the system design phase, this paper design the smart home system based on phrase speech and control commands recognition.