Trajectory tracking Optimization Control of Mecanum-Wheel Mobile Robot Based on Incremental Q-Learning

This paper proposes an optimized control method based on incremental Q-learning to address the trajectory tracking problem of Mecanum-wheeled mobile robots. Building upon a nominal controller, the method introduces an incremental controller to compensate for control input and enhance performance without altering the system structure. Notably, it requires neither the system’s dynamic model nor the control gains of the nominal controller. By incorporating the dynamic characteristics of Mecanum wheels, a piecewise linear model is introduced to dynamically update the control matrix. Incremental Q-learning is applied within each subsystem to achieve online optimal control. Simulation results demonstrate that, compared with conventional control approaches, the proposed method significantly improves trajectory tracking accuracy, thereby validating its effectiveness and engineering applicability.