Abstract:

Focusing on the steering control problem of autonomous vehicle on large curvature road, this paper presented a human-like steering control model. By imitating the judgment of human driver from the road state of the preview point, two parallel fuzzy controllers were used to estimate the steering demands independently according to the yaw angle errors and lateral errors respectively. Constrained by the road curvature radius, the output of the parallel controllers was decided by the leverage principle so as to calculate the desired tunning angle of the steering wheel. Simulation results of both double-lane change and perpendicular turn illustrate the good performance which indicates that the proposed method can control the trajectory of the vehicles effectively, which ensures the driving stability and therefore can be applied to large curvature control problems.

Key words: autonomous driving, fuzzy logic, driver model, intelligent vehicle

摘要：

针对自主车辆的大曲率转向控制问题,研究拟人转向控制模型。模仿驾驶员对预瞄点运动状态的判断,采用2个并联的模糊控制器,分别估计预瞄偏差角和偏差距离对转向意愿的判断。以道路曲率半径作为约束,利用杠杆原理对并联控制器的输出进行决策,获得期望的方向盘转角。双移线和直角弯2种道路仿真结果表明：基于预瞄最优曲率模型的大曲率转向控制方法能有效地控制车辆轨迹,运动过程平稳,适用于大曲率转向控制。


关键词: 自主驾驶, 模糊逻辑, 驾驶员模型, 智能车辆