An intelligent authentication & intrusion detection system for securing advanced metering infrastructure

Advanced Metering Infrastructure (AMI) forms a crucial part of smart grids by enabling real-time monitoring and two-way communication between consumers and utility providers. However, its reliance on wireless and multi-hop communication exposes it to spoofing and identity-based attacks. In this paper, we propose an Intelligent Authentication & Intrusion Detection System (IAIDS) for securing AMI, designed to authenticate smart meters (SMs) and continuously monitor the network to detect identity based attacks in real time. It’s a novel three-phase framework that leverages the spatial correlation of RSSI patterns and machine learning to detect identity attacks. First, IAIDS estimates the expected RSSI (ERSSI) values by leveraging physical and environmental models (weather and terrain), thereby eliminating the need for predefined signal profiles. This Phase effectively reduces false alerts by accounting for normal RSSI fluctuations caused by environmental variability. Second, it utilizes unsupervised outlier detection techniques to dynamically identify potential anomalies at the local level, without relying on static thresholds. This enables each SM to adaptively detect suspicious behavior, improving detection sensitivity. Third, IAIDS confirms the anomalies through cooperative classification among neighboring SMs, enhancing precision. Evaluated on real RSSI and weather datasets, IAIDS achieves high precision and recall. This approach ensures high detection sensitivity while maintaining a low false alarm rate, making it suitable for secure, reliable, and scalable deployments within AMI network.