Learning-Based Security Control of Unmanned Aerial Vehicle Swarm System With Multisource Disturbances and Cyber...

Area of Science:

• Robotics and Control Systems
• Artificial Intelligence
• Cyber-Physical Systems

Background:

• Unmanned Aerial Vehicle (UAV) swarms face complex challenges including multisource disturbances and cyber attacks.
• Ensuring reliable communication, connection, and collision avoidance (GCCA) is critical for UAV swarm operations.

Purpose of the Study:

• To develop a robust control strategy for UAV swarms that guarantees communication, connection, and collision avoidance (GCCA).
• To address the impact of multisource disturbances and cyber attacks on UAV swarm control.
• To achieve finite horizon H-infinity consensus performance for UAV swarms.

Main Methods:

• An improved artificial potential field (APF) function was developed for communication, connection, and collision avoidance (CA).
• A reinforcement learning (RL) model-free attitude controller was proposed, integrated with a disturbance suppression trajectory controller.
• The improved reciprocally convex combination approach was utilized to establish a consensus performance criterion.

Main Results:

• The proposed RL-based controller successfully accommodated multisource disturbances (matched and unmatched) and cyber attacks.
• The GCCA objectives were effectively achieved, ensuring stable UAV swarm operation.
• A sufficient criterion was derived to guarantee finite horizon H-infinity consensus performance.

Conclusions:

• The developed control scheme is feasible and effective for secure UAV swarm operation under challenging conditions.
• The integration of RL and GCCA provides a robust solution for complex UAV swarm control problems.
• The research contributes to the advancement of autonomous and secure multi-UAV systems.