Research And Implementation Of Embedded Automotive Electronic Equipment Control Terminal Based On Speaker-Independent Speech Recognition

With the rapidly developing of automobile electronics technology, modern automobile accelerate progress towards the direction of intelligent, green energy, safe and comfortable, more and more vehicle devices are brought into cars, such as automotive electronics, multimedia,navigation, meter terminal. The vehicle’s performance is being improved, however, due to the increasing vehicle devices, driver’s control objects also increase dramatically, this brings a lot of traffic safety problems. The developing of speech recognition and embedded technology make voice commands instead of part of operation of vehicle devices’ manual manipulation as possible. Drivers can control the vehicle devices by voice commands, combined with the former vehicle screen terminal in real time, visually see the result of operation. Thereby it reduces the burden of driving, while simply, safely completes the control of vehicle devices, it also greatly enhances the driving comfort.In this dissertation, the vehicle devices are controlled objects, a design scheme of embedded speaker-independent local voice control terminal of in-vehicle devices is presented, it’s based on HMM algorithm. The embedded processor S3C6410A with ARM1176JZF-S core is used as controller of in-vehicle devices, speech recognition module is based on LD3320, and Innolux’s10.1 inch TFT-LCD is used as vehicle screen terminal, using the development platform of Linux2.6.38 and Qt4.6. After the construction of a system of hardware and software platform, a prototype of the speaker-independent voice control terminal system of in-vehicle devices is designed and implemented, the prototype realizes the recognition of voice command, the real-time control of air-condition, audio, vehicle’s lights, vehicle’s windows, windshield wipers and the display of operating results on system terminal screen. It also can be controlled manually, and manual precedence over voice, to provide protection for the driver to safety. To improve the reliability of the system, we make the appropriate anti optimized design for the vehicle environment. Finally, performance test experiments are done on the terminal prototype system, the experimental results show that the system has many advantages, such as high recognition rate, good stability, accurately manipulation and display, friendly display interface and good scalability.