Adaptive neural network output feedback control for stochastic nonlinear systems with unknown dead-zone and unmodeled...

Area of Science:

• Control Theory
• Nonlinear Systems
• Stochastic Systems

Background:

• Stochastic nonlinear strict-feedback systems present control challenges due to unknown nonlinearities, dead-zones, and unmodeled dynamics.
• Lack of direct state variable measurements complicates traditional control design.

Purpose of the Study:

• To develop an adaptive neural network (NN) output feedback control scheme for stochastic nonlinear strict-feedback systems.
• To address challenges including unknown nonlinear uncertainties, dead-zones, and unmodeled dynamics.

Main Methods:

• Utilizing neural networks (NNs) to approximate unknown nonlinear functions.
• Designing an NN state observer to estimate unmeasured states.
• Employing backstepping design and a stochastic small-gain theorem for control scheme development.

Main Results:

• The proposed control scheme ensures that all closed-loop system variables are input-state-practically stable in probability.
• The system demonstrates robustness against unmodeled dynamics.
• Observer errors and system output are regulated to a small neighborhood of the origin.

Conclusions:

• The developed adaptive NN output feedback control is effective for the considered class of systems.
• The approach provides a robust solution for controlling complex stochastic nonlinear systems with uncertainties.
• Simulation results validate the proposed control strategy's effectiveness.