Event-triggered data-driven control of discrete-time nonlinear systems with unknown disturbance

Area of Science:

• Control Theory
• Nonlinear Systems
• Data-Driven Methods

Background:

• Nonlinear systems often face challenges with unknown disturbances, impacting control performance.
• Traditional control methods may struggle with model uncertainties and external disturbances.
• Data-driven approaches offer a promising alternative for controlling complex systems without explicit models.

Purpose of the Study:

• To develop an event-triggered, data-driven control strategy for nonlinear systems with unknown disturbances.
• To enhance system performance by reconstructing and compensating for unknown disturbances.
• To ensure the stability and boundedness of the tracking error system.

Main Methods:

• Utilizing an event-triggered mechanism to optimize control updates.
• Employing a model-free iterative learning approach for control design.
• Implementing an extended state observer (ESO) to estimate unknown system disturbances.
• Constructing a control strategy based on system input, output, and reconstructed disturbance.

Main Results:

• The proposed event-triggered, model-free iterative learning control strategy effectively manages nonlinear systems with unknown disturbances.
• The extended state observer successfully reconstructs the system's unknown disturbance.
• Sufficient conditions were established to guarantee uniform ultimate boundedness of the tracking error.
• Simulation examples demonstrated the practical effectiveness of the developed control scheme.

Conclusions:

• The presented event-triggered, data-driven control method offers a robust solution for nonlinear systems with unknown disturbances.
• The integration of model-free iterative learning and extended state observers provides a powerful framework for disturbance rejection.
• The proposed approach ensures system stability and improves tracking accuracy, validated through simulations.