Research On An Embedded Speech Recognition System Based On Neural Network

Speech recognition is the technology which can auto convert the voice signal into text or commands by specific program. In recent years, it is benefit from the rapid development of computer hardware and communication network, the research of speech recognition has made many encouraging results and the speech recognition products are already available in the market. With the development of computer storage technology, people can focus on how to improve the accuracy of speech recognition system better. Up to now, most speech recognition is based on conventional linear system theory. Actually, speech signal is a complex nonlinear process. In the actual environment, there are some limitations in the speech system based on linear system theory. Aim at improving the accuracy of speech recognition system and the generalization ability, some research and experiments are made in this paper.Speech recognition system including speech preprocess, characteristic parameters extraction, recognition model and speech synthesis.In this paper,the selection of recognition model is mainly focused and the BP neural network is chosen as the basic unit of the recognition model by comparing the various recognition algorithm. To solve the problem of low recognition rate of BP neural network, the theory of neural network ensemble is introduced for the recognition model. And then the K-means clustering method is used for the further improvement. In order to verify the effectiveness of the proposed method, some simulations and experiments are made in MATLAB platform and VC6.0platform respectively. And the results prove the effectiveness of this method.Finally, android operating system is chosen as the software platform for the system,and then the application development environment construction flow is detailed introduced according to a specific hardware platform and a speech recognition system is used successfully for the platform eventually...