Adaptive Actor-Critic Design-Based Integral Sliding-Mode Control for Partially Unknown Nonlinear Systems With Input...

Area of Science:

• Control Systems Engineering
• Nonlinear Dynamics
• Adaptive Control Theory

Background:

• Nonlinear systems often suffer from input disturbances and unknown dynamics, complicating control design.
• Integral sliding-mode control is a robust technique, but its performance can degrade with uncertainties.
• Adaptive dynamic programming (ADP) offers a powerful framework for optimal control in complex systems.

Purpose of the Study:

• To develop an integral sliding-mode control strategy for nonlinear systems with input disturbances and unknown nonlinear terms.
• To guarantee closed-loop stability and achieve near-optimal performance of sliding-mode dynamics using adaptive control.
• To integrate adaptive dynamic programming (ADP) with an actor-critic (AC) architecture for online optimal control learning.

Main Methods:

• Design of a neural network (NN)-based observer to approximate unknown nonlinear terms.
• Implementation of a disturbance observer to estimate time-varying input disturbances.
• Construction of a sliding-mode control law using NN approximations and disturbance estimations.
• Application of the ADP method with an AC structure for online learning of optimal control laws.
• Development of reconstructed tuning laws to ensure stability and convergence of NN weights.

Main Results:

• The proposed method effectively approximates unknown nonlinearities and estimates input disturbances.
• The designed sliding-mode control successfully eliminates disturbance effects and achieves desired sliding-mode dynamics.
• The actor-critic (AC) based ADP approach learns the optimal control policy for the sliding-mode dynamics online.
• Simulation results demonstrate the effectiveness and stability of the proposed integral sliding-mode control strategy.

Conclusions:

• The integrated approach of NN-based observers, disturbance estimation, and AC-based ADP provides a robust solution for controlling nonlinear systems with uncertainties.
• The method guarantees system stability and achieves near-optimal performance in the presence of time-varying disturbances.
• This work advances the application of adaptive dynamic programming in sliding-mode control for complex nonlinear systems.