基于RANMPC的移动机器人轨迹跟踪控制

Trajectory Tracking Control of Mobile Robots Based on RANMPC

摘要: 针对两轮移动机器人轨迹跟踪中的非完整约束、外部干扰与执行器饱和问题，本文提出一种结合参数自适应NMPC (ANMPC)与扩张状态观测器(ESO)的鲁棒控制策略。该策略通过ESO对集总扰动进行前馈补偿；利用位姿误差的高阶微分特征识别异常等级，动态调整控制参数；进而融合ANMPC与ESO输出，形成复合控制律。仿真表明，在圆轨迹上本策略跟踪精度高；在存在初始误差的复合轨迹下收敛快，轨迹突变后恢复时间仅为5.5 s。与固定参数NMPC相比，横向跟踪误差降低95.83%，航向角误差减少82.62%，且无失稳。

Abstract: This paper proposes a robust control strategy that combines parameter adaptive NMPC (ANMPC) and extended state observer (ESO) to address the issues of incomplete constraints, external disturbances, and actuator saturation in trajectory tracking of two wheeled mobile robots. This strategy uses ESO to perform feedforward compensation on the lumped disturbance; Using high-order differential features of pose errors to identify anomaly levels and dynamically adjust control parameters; Further integrate the outputs of ANMPC and ESO to form a composite control law. Simulation shows that this strategy has high tracking accuracy on circular trajectories; Under the composite trajectory with initial errors, convergence is fast, and the recovery time after trajectory mutation is only 5.5 seconds. Compared with fixed parameter NMPC, the lateral tracking error is reduced by 95.83%, the heading angle error is reduced by 82.62%, without any instability.