Stabilization of Neural-Network-Based Control Systems via Event-Triggered Control With Nonperiodic Sampled Data

Area of Science:

• Control Systems Engineering
• Artificial Neural Networks
• Systems Theory

Background:

• Neural-network-based control systems (NNBCSs) often face challenges with data transmission and stability under nonuniform sampling.
• Existing event-triggered and nonuniform sampling schemes have limitations when applied independently to NNBCSs.
• Efficient stabilization requires addressing both sampling strategies and controller design simultaneously.

Purpose of the Study:

• To develop a novel event-triggered data transmission mechanism for nonuniformly sampled NNBCSs.
• To construct a three-layer fully connected feedforward neural-network (TLFCFFNN)-based event-triggered controller.
• To ensure the asymptotical stability of the closed-loop system while minimizing controller updates.

Main Methods:

• Design of a new event-triggered data transmission mechanism utilizing nonperiodic sampled data.
• Modeling the closed-loop system as a state-delay system using a time-delay system approach.
• Application of the Lyapunov-Krasovskii functional approach to derive stability criteria.
• Formulation of sufficient conditions for controller and triggering parameter codesign via matrix inequalities.

Main Results:

• A hybrid event-triggered and nonuniform sampling scheme is proposed, enhancing NNBCS design.
• The closed-loop system's stability is guaranteed through derived criteria based on matrix inequalities.
• The controller design effectively reduces controller updates while maintaining system stability.
• Numerical examples validate the effectiveness and benefits of the proposed event-triggered control strategy.

Conclusions:

• The proposed event-triggered stabilization method offers a robust solution for nonuniformly sampled NNBCSs.
• The codesign approach ensures system stability and optimizes controller performance.
• This research contributes to more efficient and stable neural-network-based control systems.