Adaptive Neural Network Prescribed Performance Bounded- H_∞ Tracking Control for a Class of Stochastic Nonlinear...

Area of Science:

• Control Theory
• Stochastic Systems
• Nonlinear Dynamics

Background:

• Backstepping design for stochastic nonlinear systems typically achieves boundedness in probability but struggles with H∞ performance criteria.
• A positive constant term in stability analysis complicates achieving strict H∞ performance.
• Existing methods often assume square-integrable neural network approximation errors, limiting applicability.

Purpose of the Study:

• To develop a design strategy for the prescribed performance H∞ tracking control problem in strict-feedback stochastic nonlinear systems.
• To introduce a novel concept of bounded-H∞ performance to overcome existing design challenges.
• To design an adaptive neural network controller that guarantees prescribed performance and H∞ disturbance attenuation.

Main Methods:

• Utilized a novel bounded-H∞ performance concept.
• Employed adaptive neural network techniques to approximate unknown nonlinear functions.
• Applied Gronwall inequality for stability analysis, eliminating the need for square-integrable approximation errors.

Main Results:

• Designed an adaptive neural network prescribed performance bounded-H∞ tracking controller.
• Guaranteed all closed-loop signals are bounded in probability.
• Ensured tracking error converges to a predefined neighborhood and achieved specified H∞ disturbance attenuation.

Conclusions:

• The proposed bounded-H∞ performance concept effectively addresses the limitations of traditional backstepping methods for stochastic nonlinear systems.
• The developed adaptive neural network controller ensures robust tracking performance and disturbance rejection.
• Simulation results validate the effectiveness and feasibility of the proposed control strategy.