Adaptive Neural Tracking Control of Switched Stochastic Pure-Feedback Nonlinear Systems With Unknown Bouc-Wen...

Area of Science:

• Control Systems Engineering
• Nonlinear Dynamics
• Machine Learning Applications

Background:

• Hysteresis is a significant challenge in nonlinear switched systems.
• Bouc-Wen hysteresis models with unknown parameters and directions require advanced control strategies.
• Adaptive neural networks (NNs) offer solutions for complex system uncertainties.

Purpose of the Study:

• To develop an adaptive NN tracking control scheme for switched stochastic nonlinear pure-feedback systems.
• To address unknown direction hysteresis using Nussbaum functions and backstepping.
• To manage computational load with an optimized adaptation method.

Main Methods:

• Utilized the mean value theorem to handle pure-feedback structure challenges.
• Employed Nussbaum functions and backstepping techniques for unknown hysteresis.
• Synthesized radial basis function NNs and Lyapunov functions for control design.
• Implemented an optimized adaptation method to reduce computational burden.

Main Results:

• The proposed adaptive NN controller guarantees semiglobal uniform ultimate boundedness in probability for all closed-loop system signals.
• Tracking errors converge to a small neighborhood around the origin under arbitrary deterministic switching.
• Simulation examples validate the effectiveness of the developed control approach.

Conclusions:

• The presented adaptive NN control scheme effectively handles switched stochastic nonlinear pure-feedback systems with unknown hysteresis.
• The control strategy ensures robust performance and accurate tracking.
• The method offers a viable solution for complex control problems in nonlinear systems.