The Research Of Speech Recognition Method With Fuzzy Chaotic Neural Network

With the micro-computer, communications technology and the development of artificial intelligence,people increasingly eager to let the machine to understand the language of natural human, which prompted the birth of speech recognition technology, and cut to a very broad prospects for development.Artificial neural network has shown it's advantages in the speech recognition, that causes it to become a research hot spot. With the progress of science and technology, people are gradually improving the existing neural network model to make them more closer to the human brain function. The fuzzy chaotic neural network has auto-adapted, self-learning, chaotic search capabilities and fuzzy reasoning ability, to a certain extent, it compared to other artificial neural networks can more closer to human brain's information processing method. Therefore, this paper will study the applications of fuzzy chaotic neural network in speech recognition.First,this paper analyzes the development status of the speech recognition technology, bases on the tendency of current speech recognition research determines the direction of this paper; Then systematicly introduces basal acoustics theory knowledge and speech recognition basic principle, provides a theoretical basis to later study in this paper; Next, introduces fuzzy theory, chaos theory, chaotic neural networks,fuzzy network model, training algorithm and the modeling method for speech recognition; Finally, on the base of previous research make the simulation experiments to this speech recognition method which based on fuzzy chaos neural network, The experimental result indicated that because its fuzzy reasoning ability and the chaos characteristic, the fuzzy chaos neural network applies in the speech recognition can not only effectivly enhancement recognition rate, and can sharpen the anti-chirp ability.