Robot Trajectory Tracking Control Based on Neural Networks and Sliding Mode Control

In recent years, robot technology has evolved rapidly, becoming a key component of industrial automation. Robot control technology is one of the cores of robot technology, which improves the accuracy and flexibility. Sliding mode control is an effective method for robot trajectory tracking control, but traditional methods suffer from problems such as parameter adjustment, switching frequency of sliding mode surface, and large oscillation amplitude. A robot trajectory tracking control method based on neural networks and sliding mode control is proposed, which analyzes the differential motion of robots under non-integrity constraints. The traditional control method converged at 4.9s and the curve was relatively fluctuating, while the research method converged at 3.6s and the converged curve was relatively smooth and stable. Compared with the existing method, the trajectory deviation of the research method was smaller than that of the reference trajectory, and the robot motion was more stable. For average errors (instantaneous optimal control method), the average errors of lateral displacement, longitudinal displacement, and displacement offset angle of the comparison method were 0.99cm, 0.62cm, and 0.0211rad, respectively. The average errors of lateral displacement, longitudinal displacement, and displacement offset of the research method were 0.78cm, 0.45cm, and 0.0181rad, respectively.The research method has good convergence speed and tracking effect in robot sliding mode control, which can improve the system control performance, increase accuracy and stability, and enhance the robustness of the robot system.