Adaptive Neural Tracking Control for Switched High-Order Stochastic Nonlinear Systems.

Area of Science:

• Control Systems Engineering
• Nonlinear Dynamics
• Stochastic Processes

Background:

• Switched high-order stochastic nonlinear systems present significant control challenges due to unknown uncertainties and complex structures.
• Existing control methods often struggle with nonstrict-feedback configurations and arbitrary switching behaviors.

Purpose of the Study:

• To design robust adaptive neural tracking controllers for a class of switched high-order stochastic nonlinear systems.
• To address challenges including unknown uncertainties, stochastic disturbances, and nonstrict-feedback structures.
• To develop controllers that ensure stability and precise output tracking.

Main Methods:

• Utilizing the approximation capabilities of neural networks.
• Employing a common stochastic Lyapunov function method.
• Incorporating an improved power integrator technique for controller design.
• Developing adaptive state feedback controllers with multiple and reduced adaptive laws.

Main Results:

• All closed-loop system signals are demonstrated to be bounded-input bounded-output (BIBO) stable in probability.
• The system output is shown to almost surely track the target trajectory within a specified bounded error.
• A simplified controller with only two adaptive laws is proposed to mitigate over-parameterization issues.

Conclusions:

• The proposed adaptive neural tracking control strategies are effective for the considered class of systems.
• The developed controllers guarantee robust stability and accurate tracking performance in the presence of uncertainties and disturbances.
• Simulation results validate the practical applicability and effectiveness of the presented control approaches.