Intelligent Robot Trajectory Tracking Control Combining Mamdani Theory and Q-Learning Algorithm

This paper aims to develop an intelligent robot trajectory tracking control method with low computational complexity, good generalization ability, and an adaptive parameter adjustment mechanism. Aiming at the limitations of traditional control algorithms in complex environment adaptability and system robustness, a novel trajectory tracking control strategy is designed by combining Mamdani fuzzy theory and the Q-learning algorithm. The research results showed that through adaptive parameter adjustment and reward mechanism, the accuracy rate of the improved Q-learning algorithm reached 98%, the recall rate reached 90%, and converged at the fifth iteration, showing excellent convergence performance. The improved Q-learning algorithm reduced the number of steps in path planning and improved path fluency and efficiency while reducing the control time from 789 ms to 340 ms and improving the fitting accuracy to more than 96%. The research provides a new strategy for intelligent robot trajectory tracking control and promotes its application in a broader range of fields.