关键词: 汽车工程；辅助驾驶；驾驶意图；车道保持；人机协同

Abstract: Lane keeping assistant system is of great significance to vehicle lateral safety. However, there are many problems such as human-machine conflict in actual use. This paper proposes a lane keeping assistant control strategy under human-machine cooperation. Based on the theory of learning vector quantization neural network, the driver's driving intention identification model is constructed. The lane keeping assistant control system framework of human-machine cooperation is designed, determining the intervention conditions of the assistant system based on the deviation early warning model, using optimal curvature driver model of single point preview to calculate the optimal steering wheel angle, designing the fuzzy control system to realize the real-time calculation of co-driving coefficient. The test is carried out by the driver in the loop platform. The results show that the lane keeping assistant control strategy proposed in this paper could accurately identify the driving intention and provide a reasonable assisted machine corner according to the vehicle's deviation status to realize the lane under human-machine cooperative driving.

Key words: Automotive engineering; Assisted driving; Driving intention; Lane keeping; Humanmachine coordination