Experimental validation of an improved adaptive predefined time SMC for servo nonlinear system under multiple uncertainties

Accurate position tracking in seeker servos, robotics, and numerous industrial dynamic systems is a complex task, especially when there are variations in the parameters or any form of external disturbances. This study introduces a new Improved Adaptive Predefined Time Sliding Mode Control (IAPDTSMC) method, incorporating Extreme Learning Machine (ELM) technology, specifically tailored for second-order nonlinear systems (NLSs). The IAPDTSMC framework is comprised of three major parts: the formulation of a new predefined time sliding surface based on a novel prescribed performance function, which incorporates a tangent hyperbolic function; the design of an adaptive control law based on this modified PDT surface; and the establishment of a comprehensive control law, integrating an improved ELM with an updated output weight for disturbance compensation. The proposed advanced ELM model vigorously adjusts its output weights, directed by an adaptive control method within the Lyapunov stability paradigm, in contrast to traditional ELM, which uses pattern classification. IAPDTSMC also provides better resilience and faster convergence for the closed-loop control, significantly decreasing reliance on complex system dynamics. Finally, we have verified the performance of the proposed control law through simulation and experimental validation and provided a comparison analysis. The results demonstrate its remarkable accuracy, achieving high precision and faster RMS error convergence, with errors of less than 0.005 rad under a maximum torque loading of 0.065 N. The proposed control law exhibits significant effectiveness, robustness, and faster error convergence, reaching up to 10−3, outperforming as compared to existing techniques.