Decentralized Adaptive Neural Output-Feedback DSC for Switched Large-Scale Nonlinear Systems

Area of Science:

• Control Systems Engineering
• Nonlinear Dynamics
• Artificial Intelligence in Control

Background:

• Switched large-scale uncertain nonlinear systems present significant control challenges due to unknown parameters and unmeasurable states.
• Existing control methods like backstepping suffer from complexity explosion and require individual subsystem coordinate transformations.
• Decentralized adaptive control is crucial for managing complex, interconnected systems.

Purpose of the Study:

• To develop a decentralized adaptive neural output-feedback control strategy for switched large-scale uncertain nonlinear systems.
• To address challenges posed by unknown control coefficients and unmeasurable states.
• To extend Dynamic Surface Control (DSC) to switched systems while mitigating complexity.

Main Methods:

• A switched-dynamic-surface-based approach utilizing switched first-order filters to prevent complexity explosion.
• Dual common coordinates transformation to simplify subsystem design, avoiding recursive backstepping transformations.
• Integration of Nussbaum-type functions for unknown control coefficients and switched neural network observers for state estimation.
• Application of the average dwell time method combined with backstepping and DSC.

Main Results:

• The proposed control approach guarantees semiglobal uniformly ultimately boundedness for all closed-loop system signals.
• Tracking errors are confined to a small neighborhood of the origin under average dwell time switching signals.
• The effectiveness of the method is demonstrated through simulations on a two-inverted pendulums system.

Conclusions:

• The developed decentralized adaptive neural output-feedback control is effective for switched large-scale uncertain nonlinear systems.
• The approach successfully handles unknown parameters, unmeasurable states, and system switching.
• This work offers a robust solution for complex control problems in uncertain dynamic environments.