智能汽车换道避障路径规划与跟踪控制研究

为了提高智能汽车行驶安全性,研究了智能汽车换道避障路径规划与跟踪控制问题。在路径规划方面,给出了换道避障决策过程,提出了等速偏移函数与正弦函数加权叠加的路径规划方法,经验证此路径满足曲率约束条件;建立了车辆运动学和动力学模型,使用位姿误差方程求解了期望横摆角速度;在路径跟踪方面,将RBF神经网络与滑膜控制结合,提出了神经滑膜控制器;经仿真验证,相比于传统滑膜控制器,神经滑膜控制器不仅减弱了抖振现象,而且对路径跟踪的纵向偏差降低了200%,方向偏差降低了300%,且神经滑膜控制器鲁棒性很好。

Study on Path Planning and Tracking Control of Intelligent Vehicle Obstacle Avoidance by Changing Lanes

To improve driving safety of intelligent vehicle, path planning and tracking of intelligent vehicle obstacle avoidance by changing lanes are studied. In the matter of path planning, decision-making process of obstacle avoidance is given. Weighted starching of constant velocity offset function and sine function is put forward to plan the path, and the path satisfies curvature requirement clarified by theory and living example. Vehicle kinematical and dynamics equations are built, and expect yaw velocity is calculated by pose error equation. In the matter of path tracking, RBF neutral network and sliding mode are integrated, so that neutral sliding mode is raised. It is clarified by simulation, compared with traditional sliding mode controller, neutral sliding mode controller weakens chattering amplitude. Besides, compared to traditional sliding mode controller, vertical deviation decreases by 200%, directivity deviation decreases by 300%, and robust of neutral sliding mode controller is good.