Federated Learning-Based Intrusion DetectionFramework for Enhancing Security in Internet ofThings Environments

The exponential expansion of the Internet of Things (IoT) has introduced substantial cybersecurity challenges, rendering interconnected devices increasingly vulnerable to cyberattacks. To address these issues, we propose a federated learning-based intrusion detection framework leveraging Long Short-Term Memory (LSTM) networks and the Federated Averaging (FedAvg) algorithm. The system was evaluated on two benchmark datasets: UNSW-NB15 and ToN-IoT. Results show that the federated LSTM achieves 98.44% accuracy, F1-score of 92.09%, and a false positive rate of 0.0074, compared to the centralized model with 96.21% accuracy, F1-score of 84.10%, and a false positive rate of 0.03. These findings confirm that federated learning enhances both detection performance and privacy preservation, making it suitable for large-scale, heterogeneous IoT environments.