Decentralized sliding mode adaptive controller design based on fuzzy neural networks for interconnected uncertain...

Area of Science:

• Control Systems Engineering
• Artificial Intelligence
• Large-Scale Systems Analysis

Background:

• Sliding mode control (SMC) is effective for systems with uncertainties but suffers from control input chattering.
• Traditional SMC requires prior knowledge of the bounds of uncertainties and disturbances, limiting its applicability.
• Large-scale systems present unique challenges due to complex interconnections and external disturbances.

Purpose of the Study:

• To develop a novel controller, the fuzzy neural networks sliding mode controller (FNNSMC), for large-scale systems.
• To overcome the limitations of conventional SMC, specifically control input chattering and the need for known uncertainty bounds.
• To enhance the robustness of controllers for systems with unknown high-order interconnections and disturbances.

Main Methods:

• Integration of fuzzy neural networks (FNNs) with sliding mode control (SMC) principles.
• Utilizing the continuous output of FNNs to replace the discontinuous sign function in SMC, thereby mitigating chattering.
• Designing a controller that does not necessitate prior knowledge of the bounds of system uncertainties and disturbances.

Main Results:

• The proposed FNNSMC effectively eliminates control input chattering.
• The FNNSMC demonstrates robustness without requiring prior knowledge of uncertainty bounds.
• Simulation results confirm that the FNNSMC offers superior robustness compared to traditional SMC.

Conclusions:

• The FNNSMC is a viable and robust control strategy for large-scale systems with unknown uncertainties.
• This approach offers a significant improvement over conventional SMC by addressing chattering and simplifying design requirements.
• The FNNSMC provides a more practical and effective solution for controlling complex systems in the presence of significant disturbances.