非完整约束AGV轨迹跟踪的非线性预测控制

Nonlinear Model Predictive Control of Trajectory Tracking for Nonholonomic AGV

针对存在非完整约束和控制输入约束的三轮AGV模型,研究其非线性模型预测控制策略,提出了一种跟踪与镇定统一控制算法。首先由模型预测控制原理产生一个优化控制器,接着设计终端控制器及不变终端域来保证系统的镇定,然后设计状态观测器以进一步提高跟踪精度,最后设计避障控制模块来完善整个控制器功能。该控制算法具有一般性,对所有AGV及其他轮式移动机器人的运动学模型都适用。计算机仿真结果验证了该控制算法的正确性和有效性。

Considering the kinematic model of three-wheeled AGV with non-holonomic constraint and input saturation,a nonlinear model predictive control was studied and an united control algorithm of tracking and stability was proposed.Firstly,an optimal controller was obtained based on the principles of model predictive control.Secondly,a terminal controller and an invariantterminal region were designed in order to keep the AGV system stable.Then,the noise problem of state estimation was solved by using a state observer,which improves the localization precision and enhances the control effectiveness.Finally,a module was designed to avoid obstacle.The designed control algorithm has generality for kinematic model of other wheeled mobile robots.Computer simulation results show its correctness and validness.