Distributed Consensus of Feedforward Nonlinear Stochastic Multiagent Systems Subject to Actuator Attacks

Area of Science:

• Control Theory
• Systems Engineering
• Robotics

Background:

• Investigates distributed leader-following full-state consensus in nonlinear stochastic multi-agent systems (MAS).
• Focuses on challenges posed by actuator deception attacks in fixed directed network topologies.
• Addresses limitations of existing attack models by introducing a novel, stealthier actuator deception attack model.

Purpose of the Study:

• To develop a novel distributed output feedback consensus control algorithm for MAS under actuator deception attacks.
• To overcome the impact of sophisticated, stealthy actuator deception attacks and inherent stochastic disturbances.
• To ensure leader-following full-state bounded consensus in probability for all agents in the system.

Main Methods:

• Introduced a novel actuator deception attack model using a stochastic inverse dynamics system.
• Designed a distributed linear controller with a compensator for each follower based on relevant outputs.
• Constructed a new Lyapunov function and employed the exchanging supply function approach for rigorous proof.

Main Results:

• Developed a novel distributed output feedback consensus control algorithm effective against stealthy actuator deception attacks.
• Sufficient conditions for the controller's effectiveness were established.
• Proved that all agents achieve leader-following full-state bounded consensus in probability.

Conclusions:

• The proposed control strategy effectively mitigates the impact of stealthy actuator deception attacks in nonlinear stochastic MAS.
• The methodology ensures robust leader-following full-state bounded consensus under challenging conditions.
• Simulation results validate the efficacy of the developed control approach.