Discrete-time robust event-triggered actuator fault-tolerant control based on adaptive networks and reinforcement...

Area of Science:

• Control Systems Engineering
• Nonlinear Systems Analysis
• Fault Diagnosis and Tolerance

Background:

• Discrete-time systems are susceptible to nonlinear uncertainties and actuator faults, compromising performance.
• Existing fault-tolerant control methods may not adequately address complex fault types or optimize data transmission.

Purpose of the Study:

• To develop an adaptive fault-tolerant control scheme for discrete-time systems with nonlinear uncertainties and actuator faults.
• To enhance system robustness and minimize data transmission through an event-triggered mechanism.
• To validate the proposed controller's effectiveness using theoretical analysis and experimental results.

Main Methods:

• Design of an adaptive controller based on a nonlinear electronic circuit to manage actuator faults (offset-biasing, sensitivity variation, dead-zone).
• Implementation of an event-triggered sliding surface mechanism for improved robustness and reduced data exchange.
• Utilization of adaptive networks (MiFRENs) trained with reinforcement learning for adaptive control.
• Theoretical analysis to guarantee signal boundedness and tracking error convergence.

Main Results:

• The proposed adaptive controller effectively handles various actuator faults and nonlinear uncertainties.
• The event-triggered mechanism significantly reduces data transmission while maintaining system performance.
• Experimental validation confirms the controller's robustness, reduced data load, and superior performance compared to existing methods.

Conclusions:

• The developed fault-tolerant control scheme offers a robust and efficient solution for discrete-time systems with complex uncertainties and faults.
• The integration of adaptive networks and event-triggered control presents a promising direction for advanced control system design.
• The findings demonstrate practical applicability and significant advantages over conventional approaches.