Study On Driver Fatigue Warning System Based On Credible Service Platform For Vehicle

With the rapid development of automobile industry and the continuous improvement of living standard in recent decades, automobile has already become an important component of modern life. However, as the increasing number of motor vehicles, the number of traffic accidents has been increasing at the same time. Driver fatigue as one of the main reason of traffic accident, the driver fatigue detection and warning technology has aroused widespread concerned in the industry. Most of the existed fatigue detection devices are designed based on local image processing chips, but the local fatigue detection method will typically have some limitations due to limited hardware processing speed and uncertainties in the vehicle environment. In this article, the system designed based on the idea of remote image processing, local data of collected image was transmitted to the remote service center via network, image processing and fatigue detection are completed by the high-speed PC in the client machine of remote server, and the result will eventually transfer back to the vehicle terminal, to warn the driver simultaneously.In this paper, the functions of driver fatigue warning system are designed and implemented on basis of the credible network service platform of vehicle, it could be considered as the extension functional of credible vehicle network service platform. The credible network service platform of vehicle is designed in the next generation of Internet services in the State Development and Reform Commission and the test equipment industry specific business:"Credible Network Service Platform of Vehicle based on IPV6". The vehicle service platform is based on the Telematics architecture, with the function of driving with GPS navigation and auxiliary security.The main work of this writing:1,In basis of a detailed study of the mechanism of fatigue driving and a variety of existed detection methods, the credible automotive service platform and its main components-Vehicle Gateway are firstly introduced.2,Set up the main hardware structure of driver fatigue warning system:video input module constructed by the IR CCD camera and USB video capture card:control module based on the vehicle gateway and alarm module formed by the CAN communication nodes and the executing agencies.3,The development of device drivers for various components of the system to the hardware devices involved in LINUX environment, including video input device driver and CAN interface device driver which laid the foundation for the system correctly. Then designed video capture algorithm based on V4L and alarm program based on Microcontroller AT89S52. Finally, on the basis of the previous study in fatigue detection algorithm, a fatigue detection algorithm of three major components:face detection, eye detection and PERCLOS fatigue detection was designed, and the algorithm is achieved in Simulink environment.The innovation of the driver fatigue warning system as following:1,Views put forward by the remote image processing approaches on fatigue driving test, which meet the idea of telematics architecture and with a certain technology forward.2,Modular programming using Simulink instead of traditional complex coding with redundancy in the process of system software development, and could run simulation results at any time, which improved the efficiency of software development greatly.3,Using a combination of horizontal and vertical projection method to determine the center of the eye in eye detection algorithm. After analysis and comparison of experimental results, the algorithm is simple and easy to implement with fast processing speed and expected test results can be obtained.Due to time constraints, the system also has some aspects to be improved, such as the time delay caused by remote image transfer will reduce the real-time performance of the system and the error brought by the light intensity is not enough when night driving. In order to further improve system reliability and real-time performance, the next step is to focus on improvements and fix the problems found in experiments.