基于视觉的车道保持控制稳定性分析

Research on Stability of Vision-based Lane Keeping Controller

基于预瞄跟随理论和二自由度车辆模型,提出了一种自适应模糊PID的车道保持控制方法。将经典PID控制与专家模糊逻辑有效地结合,通过对反馈状态加入调整因子,有效并实时地为误差和误差导数合理地分配权重系数,从而实现对反馈状态的优化。用Matlab对所提出的车道保持控制方法进行了仿真,取得了较好的控制效果。与传统PID算法进行了比较,结果表明,所提出方法较传统PID控制器具有更优的动态响应性能和更佳的鲁棒性。

This paper proposes a self-adjusting fuzzy PID algorithm by taking into account of 2-DOF vehicle lateral dynamic model and preview following theory. This algorithm combines classic PID control algorithm with expert fuzzy logic together; meanwhile an adjustment factor is added to the feed- back state. The factor allocates weight coefficients for error and error＇ s derivatives in real-time to keep the feedback state optimum. At last, the control method of the lane keeping is verified by the MATLAB simulation. The comparison results with the classic PID algorithm show that the proposed self-adjusting fuzzy PID controller is of great superiority, not only has better dynamic response, but also has better robustness