Data-driven MFAC for a class of discrete-time nonlinear systems with RBFNN.

Area of Science:

• Control Systems Engineering
• Artificial Intelligence
• Nonlinear Dynamics

Background:

• Model-free adaptive control is crucial for systems where plant models are unavailable or complex.
• Discrete-time nonlinear systems present significant control challenges due to their inherent complexities.
• Neural networks offer powerful tools for online parameter tuning in adaptive control.

Purpose of the Study:

• To propose a novel model-free adaptive control (MFAC) method for discrete-time SISO nonlinear systems.
• To utilize radial basis function neural networks (RBFNN) for online controller parameter tuning.
• To ensure and verify the stability and effectiveness of the proposed control strategy.

Main Methods:

• The proposed method employs equivalent dynamic linearization of the ideal nonlinear controller.
• Radial basis function neural networks are used to tune controller parameters directly from plant input-output data.
• Rigorous theoretical analysis is performed to guarantee system stability.

Main Results:

• The model-free adaptive control method demonstrates guaranteed stability through theoretical analysis.
• Numerical simulations confirm the effectiveness and applicability of the proposed control technique.
• Experimental validation on a three-tank water level control process further substantiates the method's performance.

Conclusions:

• The developed model-free adaptive control strategy effectively manages discrete-time SISO nonlinear systems without prior model knowledge.
• The integration of RBFNN for online tuning provides a robust and practical control solution.
• The method's success in simulations and practical experiments highlights its potential for real-world applications.