Research Of Application Of Embedded Speech Recognition On Automobile GPS Navigation System

Embedded system is more and more important in both research and application fields now, and various kinds of SOC systems enter into practical application. As automobiles become popular in China, there is a great demand for GPS positioning and navigating systems and the system based on ARM best meets this demands. This system consists of a microcontroller with ARM inside and peripherals such as LCD, touch screen, USB and audio device, which is a low cost but high efficiency system.Speech recognition is one of the most convenient methods to communicate with machine, and some commercial products have achieved very good recognition result, and it has very bright future to implement the speech recognition on the embedded system. The GPS navigation system on the automobile with speech recognition ability does not need the drivers to press button anymore is also very attractive.This paper introduces firstly introduces essential principle related to the speech recognition, and explain MFCC(Mel Frequency Ceptral Coefficients),HMM(Hidden Markov Model),and VQ(Vector Quantization).Then, it comes the embedded platform, from hardware design to the software design especially the part related to the speech recognition, schematic and drivers design .At last, the details of the implementation and result have been explained.