Integral reinforcement learning based event-triggered control with input saturation.

Area of Science:

• Control Systems Engineering
• Artificial Intelligence
• Nonlinear Dynamics

Background:

• Input-saturated systems present significant control challenges.
• Traditional adaptive dynamic programming requires full system knowledge and admissible initial control.
• Computational and communication burdens limit practical applications.

Purpose of the Study:

• To develop a novel integral reinforcement learning (IRL)-based event-triggered adaptive dynamic programming scheme.
• To address control of continuous-time nonlinear systems with input saturation.
• To reduce computational and communication costs in control systems.

Main Methods:

• Integral Reinforcement Learning (IRL) for policy learning without drift dynamics knowledge.
• Single critic neural network to approximate value functions, removing the need for initial admissible control.
• Event-triggered control law design with a specific triggering threshold.
• Asymptotic stability analysis of the developed control system.

Main Results:

• The proposed IRL scheme effectively handles input-saturated nonlinear systems.
• The event-triggered approach significantly reduces computational and communication overhead.
• Guaranteed asymptotic stability of the closed-loop system is demonstrated.
• Simulation results validate the effectiveness of the developed control method.

Conclusions:

• The developed IRL-based event-triggered adaptive dynamic programming scheme is effective for input-saturated nonlinear systems.
• This method offers a practical solution for reducing control system costs while ensuring stability.
• The approach advances the application of AI in robust control engineering.