Driving State Recognizing Based On Driving Behavior Signals

At present, the method of detection driving state based on physiological and psychological is accurate. But it affected the driving operation behavior, so vehicle-aided detection system is not possible. With the continuous development of computer technology, detection technology, as well as sensor technology, these provides a tremendous motivation and development space for driving state recognizing based on driving behavior signals. The method of driving state recognizing based on driving behavior signals has characteristics such as signal acquisition convenient, fine stability, high reliability. And this approach for vehicle-aided detection system is possible. The research and detect of driving state based on driving behavior signals has a very practical significance and has abroad prospect for development.The method of driving state recognizing based on driving behavior signals was described in this paper in detail. At the first of the paper formative factor of the fatigue driving and its current development was introduced. Then driving fatigue experimental design based on driving simulation and driving behavior signal collection were describer. At the end of the paper different methods of driving state recognizing was introduced, such as wavelet entropy, wavelet singularity detection and GMM. There're individual differences exist in driving behavior, which means one driver's behave differed from another when driving. Besides, even one person has differences behave in different driving state. So, there were obviously different driving behavior in the normal and fatigue state. In this paper, using steering wheel angel and the distance to center of lane, PERCLOS80 was used as reference for the state of fatigue. We study the different method of identification of the fatigue driving state, by analyzing the driver's behavior characteristics of the normal driving state and the fatigue driving state.In this paper, we studied the method of detection driving state based on driving behavior signals. This method used a variety of sensors collecting driving signal, it has many merits such as non-contact and good ability real-time. It can effectively overcome the defects of contact, thus it adapts to the development trend of vehicle-aided detection system.