Self-triggered neural tracking control for discrete-time nonlinear systems via adaptive critic learning.

Area of Science:

• Control Systems Engineering
• Nonlinear System Analysis
• Artificial Intelligence in Control

Background:

• Traditional event-based control methods require continuous monitoring of triggering conditions, increasing computational load.
• Designing optimal tracking control for discrete-time nonlinear systems presents significant challenges.
• Real-time adaptation and optimal policy adjustment are crucial for advanced control applications.

Purpose of the Study:

• To develop a novel self-triggered optimal tracking control method for discrete-time nonlinear systems.
• To reduce the computational burden associated with traditional event-based control strategies.
• To enhance the triggering performance and overall control policy optimization.

Main Methods:

• An augmented plant is constructed by integrating system state and reference trajectory.
• A self-sampling function, dependent on tracking error, determines the triggering instants.
• Online action-critic learning with model, critic, and action neural networks is employed for real-time policy adjustment.

Main Results:

• The proposed method effectively reduces computational burden by eliminating continuous triggering condition evaluation.
• The self-triggered control strategy demonstrates excellent triggering performance.
• Optimal tracking control is achieved through real-time adjustment of the control policy.

Conclusions:

• The developed self-triggered optimal tracking control method provides an efficient and effective solution for discrete-time nonlinear systems.
• The approach ensures system stability while optimizing control performance.
• Experimental validation confirms the effectiveness of the online self-triggered tracking control strategy.