Design And Realization Of One-Shot Vehicular Voice-User Interface System

With the rapid development of China’s auto industry and the increase in the number of cars, Automotive electronic devices become more prosperity and more and more automotive embedded devices has added to the car. With bringing up comfort and performance to the car, these devices also increases the complexities of the drive and a hidden insecurity at the same time. Due to the limitations of the vehicle environment if we use the devices embedded the car by graphical user interface, it is more difficult and dangerous when driving. But if we use the devices through voice interactive technologies, all above problems can be solved perfect. With the voice recognition technologies, the vehicle equipments can understand human language, call up songs, radio, phones, navigation and other functions on command with voice demands; with the use of speech synthesis technology, car status, current road conditions, new telephone call and other information can broadcast directly to the driver, the driver do not need to transfer his eyes to see the information.The current on-board speech interactive system in automotive applications are limited to a small amount of vehicle electronic equipment and can not cover all the applications, such as sophisticated navigation softwares only have speech function, blue-tooth phone can only realize voice call, etc. Voice technology just like a brocade functions, can not replace graphical user interface totally. On the other hand, the current vehicle application generally uses many conversational voice identification, it do not reduce the operating time compared to the graphical user interface, and many interactions also reduce the recognition rate. Complicated voice interactions can not reflect the operation efficiency, and reduce the maneuverability of the system.Against to the special environment int the car, a One-Shot voice interaction program is proposed in this paper, the driver give a voice command to the voice interactive system and control the vehicle equipment to finish some functions like navigation, radio, bule-tooth phone and other car applications, and the car equipment also feedback some information to prompt by voice.During the interaction between the driver and the system, the driver don’t need to look the screen, and just give a voice command to manipulate the vehicle equipment, so it abandon the Complicated multi-step interactions to confirm the voice information the system broadcast. Compared to traditional interactive interface and a series of conversational speech interaction, it greatly reduces the posibility of the danger, but also to improve the operability of the vehicle equipment.Finally, we realized the One-Shot voice interactive system on the basis of iFLYTEK voice technology and WINCE5.0platform, the system is mainly composed of dictionary management module, the voice interactive service module, the speech recognition module, speech synthesis module and audio task control module. The task of dictionary management module is to compile grammar file to generate dictionaries which the Aitalk voice recognition engine needs. If you need to add voice control applications, you can just modify the syntax file to increase the corresponding syntax commands, which make the car voice interaction system has good scalability; the voice interactive service is responsible for calling speech synthesis module for voice broadcast and calling the voice recognition module to recognize driver’s voice command and at the same time,sending the recognition result to automotive applications; audio task control module’s main task is to monitor the audio task all time, and get the audio tasks change, when a new audio task applicated, according to the priority to decide whether the task should be executed to response to the user, so this module make the voice applications can work orderly to protect the voice recognition from disturbance.The system has tested on the WinCE platform, it performs more stability and easier to voice interaction. Under the environment without noise, the system recognition rate can reach95%. And when driving in the low vehicle noise and closing the window, the recognition rate can also reach more than85%, and the voice interaction time can be controlled within7seconds, so the voice technology and automotive system is realized simple and efficient union.