Resilient Cooperative Optimal Output Regulation Control for Nonlinear Multiagent Systems.

Area of Science:

• Control Systems Engineering
• Artificial Intelligence
• Cybersecurity

Background:

• Multiagent systems (MASs) are susceptible to denial-of-service (DoS) attacks, compromising their cooperative optimal output regulation (COOR).
• Existing resilient control strategies often do not adequately address nonlinear dynamics in MASs.

Purpose of the Study:

• To develop a resilient adaptive distributed observer-based control strategy for nonlinear strict-feedback MASs under DoS attacks.
• To achieve cooperative optimal output regulation (COOR) in MASs despite external disturbances and unknown system dynamics.

Main Methods:

• Construction of resilient adaptive distributed observers for leader state estimation.
• Design of a control input using feedforward and feedback components.
• Application of neural networks (NNs) and an off-policy integral reinforcement learning (IRL) algorithm with actor-critic NNs (A-C NNs) for optimal control law derivation.

Main Results:

• Successful estimation of leader dynamics and states by followers via adaptive observers.
• Development of an optimal feedback security control law robust to unknown nonlinear dynamics.
• Demonstration of the strategy's feasibility and effectiveness through numerical and practical simulations.

Conclusions:

• The proposed resilient COOR control strategy effectively handles nonlinear dynamics in MASs under DoS attacks.
• This approach significantly broadens the applicability of resilient COOR control in real-world scenarios compared to linear system limitations.