AI-Enhanced Adaptive Sliding Mode Control for Inverted Pendulum on a Cart: A Simulation Study

Abstract

The inverted pendulum on a cart is a canonical benchmark in control engineering, representing a highly nonlinear and inherently unstable system. This paper investigates the application of an AI-enhanced adaptive sliding mode control (SMC) strategy for its stabilization and control. The proposed approach cooperatively combines the robustness of SMC against uncertainties and disturbances with the adaptive capabilities of an online adaptation law and the smoothing properties of a simplified fuzzy logic system. This integration facilitates the dynamic adjustment of control gains, effectively mitigating chattering and improving overall system performance. Simulation results, obtained using MATLAB, demonstrate the effectiveness of the proposed control strategy in stabilizing the pendulum in the upright position and achieving precise cart positioning. The system's performance is thoroughly analyzed through time-domain responses, phase-plane analysis, control effort evaluation, robustness tests under parameter variations and external disturbances, and an examination of adaptive gain dynamics. A Lyapunov stability analysis is provided to prove the asymptotic stability of the closed-loop system. Comparative studies with standard SMC, boundary layer SMC, and adaptive SMC with fixed adaptation rate demonstrate the superiority of the fuzzy-adaptive approach.