Neural-Network-Based Adaptive Event-Triggered Consensus Control of Nonstrict-Feedback Nonlinear Systems.

Area of Science:

• Control Systems Engineering
• Nonlinear Dynamics
• Artificial Intelligence

Background:

• Consensus control is crucial for multi-agent systems but often requires continuous communication.
• Nonlinear systems with nonstrict-feedback present significant control challenges.
• Dynamic leaders in consensus networks add complexity to control design.

Purpose of the Study:

• To develop an event-triggered consensus control strategy for nonlinear systems with a dynamic leader.
• To approximate unknown system dynamics using neural networks (NNs).
• To reduce communication overhead in distributed control systems.

Main Methods:

• Utilized neural networks (NNs) for approximating unknown nonlinear dynamics.
• Designed a novel adaptive event-trigger condition based on relative outputs, NN weights, and follower states.
• Employed backstepping control design with adaptive NN parameter laws to overcome algebraic loops.
• Developed an adaptive NN controller for event-triggered leader-following consensus.

Main Results:

• Achieved ultimately bounded leader-following consensus without Zeno behavior.
• Significantly reduced data communication and controller update frequency.
• Overcame the algebraic loop problem inherent in NN-based control design.
• Demonstrated controller effectiveness through simulation studies.

Conclusions:

• The proposed adaptive NN event-triggered controller is effective for nonlinear systems.
• The approach enhances communication efficiency in multi-agent systems.
• Guaranteed consensus and stability are achieved with reduced communication burden.