Collision-Free Formation Control for Heterogeneous Multiagent Systems Under DoS Attacks.

Area of Science:

• Robotics
• Control Systems
• Cyber-Physical Systems

Background:

• Heterogeneous multi-agent systems face significant challenges from cyber threats like denial-of-service (DoS) attacks, noncooperative dynamic obstacles, and input saturation.
• Existing control frameworks often struggle to integrate cyber-layer resilience with physical-layer safety guarantees for complex systems.

Purpose of the Study:

• To propose a safe time-varying formation (TVF) control framework for heterogeneous multi-agent systems.
• To address the combined challenges of DoS attacks, noncooperative dynamic obstacles, and input saturation.
• To enhance system resilience and ensure operational safety.

Main Methods:

• A cyber-layer distributed resilient observer using control Lyapunov function (CLF)-quadratic program (QP) to estimate a reference exosystem and decouple dynamics.
• A physical-layer collision-free TVF controller employing CLF-exponential control barrier function-QP for safety guarantees.
• Integration of cyber-layer and physical-layer components for a comprehensive framework.

Main Results:

• The proposed observer enhances system resilience against DoS attacks by optimizing network decoupling.
• The novel TVF controller ensures collision-free operation and safety for high-order heterogeneous agents under noncooperative obstacles and input saturation.
• Comparative simulations and experiments on unmanned aerial vehicles and unmanned ground vehicles validate the framework's effectiveness.

Conclusions:

• The integrated cyber-physical framework effectively manages heterogeneous multi-agent systems under adversarial conditions.
• The developed methods provide robust safety and resilience, outperforming existing approaches.
• This work establishes a new standard for secure and safe control in complex multi-agent applications.