Fuzzy Multiple Hidden Layer Recurrent Neural Control of Nonlinear System Using Terminal Sliding-Mode Controller.

Area of Science:

• Control Systems Engineering
• Artificial Intelligence
• Nonlinear Dynamics

Background:

• Nonlinear systems present significant control challenges due to their complex dynamics.
• Existing control methods often struggle with approximation accuracy and robustness.
• Recurrent neural networks and fuzzy logic offer potential for enhanced nonlinear control.

Purpose of the Study:

• To design a novel fuzzy double hidden layer recurrent neural network (FDHLRNN) controller for nonlinear systems.
• To integrate FDHLRNN with terminal sliding-mode control (TSMC) for improved performance.
• To enhance nonlinear approximation accuracy and dynamic control capabilities.

Main Methods:

• Developed a fully regulated FDHLRNN by combining fuzzy neural networks (FNN) and radial basis function neural networks (RBF NN).
• Incorporated outer layer feedback to boost dynamic approximation ability.
• Applied FDHLRNN with TSMC to approximate the nonlinear sliding-mode equivalent control term, reducing switching gain.

Main Results:

• The proposed FDHLRNN demonstrated superior nonlinear approximation accuracy.
• FDHLRNN with TSMC achieved faster convergence speeds compared to traditional methods.
• Simulations and hardware experiments with an active power filter confirmed robust performance and feasibility.

Conclusions:

• The FDHLRNN controller offers significant advantages in accuracy and dynamic approximation for nonlinear systems.
• The integration of FDHLRNN with TSMC provides robust and efficient control solutions.
• The method is validated for practical applications, including active power filtering.