Induced attack with prescribed consensus trajectory against multiagent systems.

Area of Science:

• Control Theory
• Cybersecurity
• Robotics

Background:

• Multiagent systems (MAS) are susceptible to various attacks.
• Existing attacks like denial-of-service or deception attacks have limitations in controlling system-wide behavior.
• A new attack paradigm is needed to precisely dictate MAS trajectories.

Purpose of the Study:

• To propose a novel induced attack strategy for multiagent systems.
• To enable an attacker to prescribe a specific trajectory for the entire MAS.
• To develop a method for generating and implementing such an attack.

Main Methods:

• An attack generation exosystem is established to create the induced attack signal.
• A new induced attack protocol is designed using local state information and the attack signal.
• The robust H_∞ scheme is employed to design the attack generation exosystem dynamics.

Main Results:

• An explicit expression for the prescribed consensus trajectory is derived.
• The induced attack successfully drives the MAS to follow the attacker's desired trajectory.
• The H_∞ design criterion ensures the effectiveness of the attack strategy.

Conclusions:

• The proposed induced attack strategy is effective in controlling the global behavior of multiagent systems.
• This attack offers a new capability beyond traditional denial-of-service or deception attacks.
• The theoretical results are validated through simulation examples.