Neural Network‐Based Adaptive Super‐Twisting Control for Perturbed Nonlinear Systems: Lyapunov Design

ABSTRACT A newly developed robust control strategy based on an adaptive neural network (ANN) estimator performed in concert with adaptive super‐twisting (ASTW) control is presented. The proposed adaptive robust control algorithm is applied to a class of perturbed nonlinear systems and features a significantly decreased control gain with respect to the unknown boundary of the perturbation. Convergence to a real second‐order sliding mode (2‐SM) is provided. The proposed adaptive protocol in the ASTW approach is simplified compared to those studied in the literature, removing the use of the differentiator and reducing the number of adaptive gains and tuning parameters. A novel Lyapunov function is used for deriving updated weights of the ANN estimator and for proving a finite‐time convergence to real 2‐SM. The efficacy of the proposed algorithm is verified via simulations.