The Design Of Voice Control System In Car And Realization On DSP

Electronics and intelligence has become a popular theme of current automotive industry in the development of global automotive industry. Automobile manufactures have reached a broad consensus on the application of intelligent speech technology in car. This technology provides a new and vast value-added space to the expansion of the entire automotive industry chain. With the characteristics of low cost, small bulk, high precision and fast processing speed, DSP chips are widely used in military, transportation, office and so on. Therefore, people are paying more and more attention to the voice control system in car based on DSP.This thesis first briefly introduces the basic theory of speech recognition and summarizes the methods for robust speech recognition. Then, it elaborates the hardware design and software realization of the voice control system in detail. It also discusses how to improve the robustness of control system in car.Based on TMS320C6713DSK, three parts are respectively designed for the control system. They are input module, data processing module and output module. To ensure that the system runs program in standalone mode when it is powered up independently of the debug environment, this paper also introduces a solution to realize the boot function.The system can recognize twelve control commands of car navigation for independent speakers in TMS320C6713DSK. The recognition accuracy of this system is 98.52% in quiet environment.In addition, the thesis makes specific algorithm analysis for the control system in car. Least Mean-Squared Error Estimation of Mel-Frequency Cepstral Coefficients Using a Novel Distortion Model is adopted to improve recognition accuracy. Compared to several traditional methods, the test shows that this method has good robustness.