A novel adaptive neural network-based time-delayed estimation control for nonlinear systems subject to disturbances...

Area of Science:

• Control Systems Engineering
• Nonlinear Dynamics
• Artificial Intelligence

Background:

• High-order nonlinear systems (HNSs) often face challenges with unknown dynamics and external disturbances, hindering precise tracking performance.
• Traditional model-based control methods require accurate system models, which are often unavailable or difficult to obtain for complex systems.
• Existing model-free control strategies may struggle with the complexity and stability of high-order nonlinear systems.

Purpose of the Study:

• To develop an adaptive backstepping-based model-free control (BSMFC) strategy for general high-order nonlinear systems.
• To enhance the tracking performance of HNSs in the presence of disturbances and unstructured uncertainties.
• To address the design complexity and approximation errors inherent in model-free control techniques.

Main Methods:

• The proposed BSMFC integrates backstepping control (BSC) with radial basis function neural network (RBFNN)-based time-delayed estimation (TDE) to handle unknown system dynamics.
• A command-filtered (CF) approach is employed to mitigate the complexity explosion typically associated with BSC design.
• Novel control laws are formulated to minimize approximation errors, ensuring improved control accuracy.

Main Results:

• The adaptive BSMFC methodology demonstrates effective control for HNSs with unknown dynamics and disturbances.
• The integration of RBFNN-TDE and CF techniques successfully overcomes the limitations of traditional model-free approaches.
• Simulation results confirm the stability of the closed-loop system via Lyapunov theorem and highlight the superiority of the proposed method.

Conclusions:

• The developed adaptive BSMFC provides a robust and effective solution for controlling high-order nonlinear systems without explicit system models.
• The combination of advanced control techniques significantly enhances system tracking performance and stability.
• This approach offers a promising direction for model-free control applications in complex dynamic environments.