基于前向道路信息的拖车倒车系统滚动时域控制

摘要

针对离轴联接形式的拖车倒车系统,阐述了其混合逻辑(mixed logical dynamics,MLD)的建模理论和基于滚动时域控制(receding horizon control,RHC)的方法.首先,采用Lyapunov理论分析了拖车倒车系统的运动学特性.接着,采用MLD建模理论来描述拖车倒车系统的运动学模型.针对此MLD运动学模型,提出了一种扩展的二次型代价函数.为了减小拖车的跟踪误差以及平滑牵引车转向角的操控性能,提出将其等效为线性二次型跟踪问题.针对此问题,利用多参数混合整数二次规划(multiparametric mixed-integer quadratic programming,mp-MIQP)技术,设计了一种考虑有限时域前向道路信息的显式RHC控制器.最后,通过仿真对控制器进行验证.结果表明尽管跟踪非平滑的参考路径,控制器均表现出期望的控制效果.%On account of the backing-up tractor-trailer system with an off-axle connection,a mixed logical dynamics (MLD) approach and a receding horizon control (RHC) method have been addressed respectively for the purpose of modeling and control of such a system.At first,Lyapunov theory was used to analyze the motion characteristics of the tractor-trailer system.Then,MLD approach was proposed to describe the kinematics of the tractor-trailer system.For the resulting MLD model,an extended cost function with quadratic form was addressed.To minimize the tracking error of the trailer,and smooth the behavior of the steering angle,an equivalent linear quadratic tracking problem was presented.In view of this problem,an explicit RHC controller was designed considering a look-ahead path information with finite horizon,by employing the multi-parametric mixed-integer quadratic programming (mp-MIQP) technique.Finally,the controller was evaluated through simulation,achieving a desirable performance,even under the tracking of a non-smooth reference path.