A systematic review of anomaly detection in IoT security: towards quantum machine learning approach

Abstract

Integrating IoT into daily life generates massive data, enabling smart factories and driving advancements in related technologies like cloud/edge computing, ML, and AI. While ML has been used for data analysis and forecasting, challenges such as data complexity, security, and computing limitations persist, particularly in anomaly detection crucial for network security. Recent research indicates the potential of quantum computing and Quantum Machine Learning (QML) to outperform traditional methods in anomaly detection within IoT, an area lacking a comprehensive review. This paper presents a systematic review of Machine Learning-based anomaly detection techniques for IoT security. Despite previous reviews, this study includes the analysis of feature engineering and quantum machine learning techniques in literature. Our findings show that current models have high detection rates on known datasets, but face scalability, real-time processing, and generalization issues. Privacy and security concerns in federated learning (FL) and the effects of data drift also need to be addressed, along with the challenges of 5G and 6G-enabled IoT environments. Future directions include integrating Explainable AI into anomaly detection, exploring adaptive learning techniques, and combining blockchain with machine learning models. The study also highlights the potential of quantum computing to enhance threat detection through quantum machine learning models.