FedHNN: A Federated Learning Based Hybrid Neural Network for Real-Time Intrusion Detection Systems

Wireless Sensor Networks are vulnerable to intrusions. Intrusions are increasing with demand of Wireless networks creating threats of information leaks and disruption. Centralized anomaly based intrusion detection systems creates bottleneck in bandwidth and increase energy utilization of the nodes. In this paper, a distributed learning approach is employed to update the intrusion detection system in real-time. A federated learning based CNN-LSTM is trained on several local client nodes for real-time updates of the system. Distributed training and centralized testing is performed on the NSL-KDD dataset. The proposed gave high accuracy of 97.68% with very low loss of 0.1568. The results outperform state of the art methods with low computational requirements.