To address the degradation in path tracking accuracy caused by conflicts between speed maintenance and turning capability in differential-drive robots operating in complex environments, a combined control strategy integrating active speed adjustment and adaptive-preview Linear Model Predictive Control (AP-LMPC) is proposed. Based on the kinematic model of the robot, a coupling constraint equation between path curvature and feasible longitudinal velocity is first derived, and an active speed adjustment strategy tailored to curvature features is designed. Then, a mapping between driving speed and optimal preview distance is constructed to enable the adaptive-preview mechanism. These components are integrated into an AP-LMPC-based control system to enable adaptive path tracking with active speed regulation. Experimental results demonstrate that the proposed strategy can significantly reduce displacement and heading errors. Specifically, the peak displacement and heading errors are maintained within 0.05 m and 0.09 rad under complex S-shaped and double-lane-change trajectories. Compared with fixed-preview LMPC (FP-LMPC) and nonlinear MPC (NMPC), the displacement error is reduced by over 68% and 48%, respectively.

differential-drive robot, path tracking, linear model predictive control (LMPC), active speed adjustment; adaptive-preview

针对复杂环境下差动机器人因速度保持与转向能力存在冲突而导致路径跟踪精度下降的问题, 提出一种融合主动速度调节与自适应预瞄线性模型预测控制(AP-LMPC)方法。基于差动机器人运动学模型, 首先推导路径曲率与可行纵向速度的耦合约束方程, 设计面向路径曲率特性的主动调速策略; 然后构建行驶速度与最优预瞄距离之间的映射关系, 实现预瞄距离的自适应调节; 最终构建基于AP-LMPC的主动调速路径跟踪系统。实验结果表明, 主动调速策略可以显著地降低路径跟踪过程中的位移和航向误差; 结合自适应预瞄机制后, 系统在S形路径和双移线路径下的位移误差峰值和航向误差峰值分别控制在0.05 m和0.09 rad内, 与固定预瞄线性模型预测控制(FP-LMPC)和传统非线性模型预测控制(NMPC)方法相比, 位移误差峰值分别降低68%以上和48%以上。