视觉导航智能车辆横向运动的自适应预瞄控制

Adaptive Preview Control of Vision Guided Intelligent Vehicle Lateral Movement

车道保持系统中车辆横向运动控制应模拟驾驶员的横向操纵行为,驾驶员根据前方道路曲率及车辆速度适时调节预瞄距离,以获得理想的路径跟踪性能。首先,以车辆二自由度动力学模型及车辆道路几何位置关系为基础,建立车-路横向动力学模型。其次,基于单点预瞄最优曲率模型设计侧向加速度PD跟踪控制器,联立车-路横向动力学模型构建横向控制闭环系统,分析预瞄距离、车速、道路曲率的变化对系统响应的影响。最后,设计模糊控制器对预瞄距离进行模糊选择以提高车辆横向控制精度和减小侧向加速度,采用遗传算法对模糊规则进行优化以使横向控制系统性能达到最优。试验表明,相对固定预瞄控制方法,自适应预瞄减小了车辆侧向加速度,且道路跟踪的方向偏差和距离偏差均得到减小。

Lateral control of vehicle for lane keeping system should simulate human driver direction control behavior, driver adaptively tuned look-ahead distance according to longitudinal velocity and road curve to obtain desired path following performance. First, a vehicle-road lateral dynamics model was established based on two-degree-of-freedom vehicle model and vehicle-road kinematics. Then a PD controller was designed to track the desired lateral acceleration based on ＂single point preview optimal curvature model＂, a lateral closed loop control system was established, which included vehicle-road lateral dynamics and PD controller, the effects of look-ahead distance, longitudinal velocity and road curve on closed loop control system responses were discussed. Finally, a fuzzy logic controller was designed to choose look-ahead distance based on fuzzy rules so as to make good lane keeping performance and small lateral acceleration. At the same time a way of optimizing fuzzy rules by using GA（genetic algorithm） was proposed to obtain optimal lateral control system performance. Experimental studies show the lateral acceleration is reduced compared with fixed point preview control method and the lane tracking deviation of direction and position decrease respectively by using adaptive preview control.