Adaptive observer-based sliding mode control with time-varying learning using Riccati gain synthesis for robotic manipulators

Abstract

This paper presents an adaptive observer-based sliding mode control (AOSMC) framework integrated with time-varying learning rates and Riccati gain synthesis (RGS) for high-precision trajectory tracking of robotic manipulators under dynamic uncertainties and external disturbances. Conventional sliding mode control methods, while robust, often suffer from high chattering and require precise knowledge of system bounds. To address these limitations, the proposed AOSMC-RGS architecture combines a nonlinear disturbance observer with Riccati-based adaptive gain tuning, enabling real-time disturbance estimation and dynamic gain adjustment based on tracking errors. A rigorous Lyapunov stability analysis ensures boundedness and convergence of system states. Simulation studies on a 2-DOF robotic manipulator demonstrate significant improvements in tracking accuracy, disturbance rejection, and control smoothness compared to PID, SMC, ASMC, Fuzzy-ASMC, and RBF-ASMC controllers. The proposed approach achieves reduced overshoot, faster settling time, and lower control effort while maintaining robustness, making it a promising candidate for real-time robotic and mechatronic system applications.