基于MPC的智能拖拉机实时路径跟踪

Path Tracking of Intelligent Tractor Based on MPC

针对具有非线性特征的智能拖拉机运动控制问题,提出了一种基于模型预测控制(MPC)的横向轨迹实时跟踪控制策略。从非线性运动模型线性化和离散化入手,建立以前轮转角增量和纵向速度增量为状态量的目标函数;设计系统控制量极限约束和控制增量约束,并将MPC最优求解问题转化为二次规划问题,目标函数采用有松弛因子和无松弛因子2种形式。基于Matlab/Simulink平台设计MPC控制器,进行了2种目标函数的固定坐标系下的给定轨迹跟踪试验。结果表明:2种控制算法均可以快速跟踪期望路径,目标函数具有松弛因子的MPC控制器的拖拉机跟踪效果和平稳性更好,前轮转角始终被限定在约束范围内。

In order to investigate the motion control with high nonlinearity for intelligent tractor,a lateral control method was proposed based on model predictive control(MPC).Although tractor model has high nonlinearity,the model selected the linear time-varying error model of kinematic model as the predictive model to improve operation velocity.Objective function that selected the control increments,longitudinal velocity increment and front wheel angle increment,as state variables was established.Meanwhile,two control schemes with and without relaxation factor were proposed.With the design of the limiting constraints of control variable and control increment,the solution of objective function was changed into quadratic programming problem.Both of model predictive controllers were designed based on Matlab/Simulink.Results showed that both of path tracking controllers effectively track the reference path and the scheme with relaxation factor had smaller displacement errors and heading errors.Meanwhile front wheel angle were always restricted within the constraint range.