Revolutionizing Security: Leveraging Unsupervised Anomaly Detection and Autoencoder Techniques to Enhance Device Security
Sujith Kumar Mandala · Zenodo (CERN European Organization for Nuclear Research) · 2023
The rapid growth of Internet of Things (IoT) devices has led to an increased need for advanced security measures to protect these devices from cyber-attacks. In this paper, we present a comparative study of unsupervised anomaly detection and autoencoder-based methods for detecting security breaches in IoT-enabled devices like number locks and fingerprint locks. We evaluate the performance of different unsupervised anomaly detection methods, such as clustering-based, density-based, and reconstruction-based approaches, in detecting security breaches. In addition, we investigate the use of autoencoder neural networks for detecting abnormal patterns in sensor data from security locks. We also explore different ways of integrating these two techniques and evaluate their performance in detecting security breaches. Our study is based on real-world data from security locks, and we evaluate the performance and limitations of these techniques in practical scenarios. Finally, we investigate different techniques to improve the robustness of these methods against adversarial attacks. Our study provides valuable insights for the field of IoT security and highlights the potential of unsupervised anomaly detection and autoencoder-based methods for detecting security breaches in IoT-enabled devices.