A two layer driver steering prediction model is proposed, in which the first layer embodies path following characteristic based on driver visual previewed information, the second layer reflects driver's steering be-haviors based on driver neuromuscular model, and the path tracking performances of driver in different states are compared by using CarSim/Simulink. The desired yaw rate observer and the steering wheel angle PID controller in LDAS are designed and the equivalent dynamics model for steering system is established, with robust torque control-ler of LDAS designed based on sliding mode theory. Due to the relationship of the degree of lane departure with the lateral offset of previewed point and driver torque can't accurately be described, the fuzzy controller for man-machine shared driving in LDAS is designed based on fuzzy control theory. A simulation with CarSim/Simulink and a hard-ware-in-the-loop test with CarSim/LabVIEW RT are conducted to compare the rectification ability of vehicle in lane departure by driver, LDAS controller and man-machine shared driving respectively. The results show that the pro-posed control strategy for man-machine shared driving can timely rectify lane departure and make vehicle back to normal lane, while ensuring smooth switching from man-machine shared control to driver control.%提出了两层驾驶员转向预测模型,基于驾驶员视觉预瞄信息的第一层体现了路径跟踪特性,基于神经肌肉动力学模型的第二层体现了驾驶员转向操作特征,采用CarSim/Simulink对比了不同状态驾驶员的路径跟踪性能.设计了车道偏离防避系统(LDAS)的期望横摆角速度观测器和转角PID控制器.建立了转向系统等效动力学模型,并基于滑模理论设计了LDAS的鲁棒转矩控制器.由于车辆偏离车道程度与预瞄点的侧向偏移量和驾驶员力矩的关系不能精确描述,故基于模糊控制理论设计了LDAS人机共驾模糊控测器.进行了基于CarSim/Simu-link的仿真和基于CarSim/LabVIEW RT的硬件在环试验,对比了驾驶员、LDAS控制器和人机共驾纠正车辆偏航的能力.结果表明,所提出的人机共驾策略能及时纠正车辆偏航,使之恢复到正常车道,并保证从人机共驾到驾驶员控制切换过程的平顺性.
展开▼
