Event-Triggered Federated Reinforcement Broad Learning for Intelligent Fault Diagnosis in Uncrewed Helicopter Swarm...

Area of Science:

• Aerospace Engineering
• Artificial Intelligence
• Control Systems

Background:

• Uncrewed helicopter swarms face complex fault diagnosis challenges due to dynamic environments and mission criticality.
• Existing methods struggle with data privacy, adaptive decision-making, and efficient resource utilization.

Purpose of the Study:

• To propose a novel Event-Triggered Federated Reinforcement Broad Learning (ET-FRBL) framework for intelligent fault diagnosis.
• To enhance the reliability, security, and efficiency of fault diagnosis in uncrewed helicopter swarm systems.
• To improve diagnostic accuracy and system performance in complex operational scenarios.

Main Methods:

• Integration of Federated Learning (FL) for data privacy and collaborative learning.
• Application of Reinforcement Learning (RL) for adaptive decision-making in dynamic conditions.
• Utilization of Broad Learning System (BLS) for rapid fault pattern recognition.
• Implementation of an event-triggered mechanism for optimized communication and computation.

Main Results:

• A Hardware-in-the-Loop (HIL) fault simulation platform was developed for comprehensive testing.
• The ET-FRBL framework demonstrated superior diagnostic accuracy compared to existing methods.
• The framework effectively handles diverse fault scenarios, including actuator and sensor failures.

Conclusions:

• The ET-FRBL framework offers a robust and efficient solution for intelligent fault diagnosis in uncrewed helicopter swarms.
• The proposed approach significantly improves diagnostic accuracy and system reliability.
• ET-FRBL addresses key challenges in data privacy, adaptivity, and resource optimization for swarm systems.