Detection of Ransomware Attack in IoT-Based Smart Meter Using Hybrid Deep Learning Model

Massive utilization of IoT-enabled consumer electronic devices such as fans, cooling and heating equipment, etc., controlled through the energy management controller interconnected with the smart meters, are highly susceptible to cyber-attacks. Any cyber-attack on the smart meter may lead to changes in the operating cycle or leakage of information from these consumer electronic devices. Protecting these consumer electronic devices from vulnerabilities necessitates an effective solution to detect vulnerabilities in the smart meter. For this purpose, in the presented study, a deep learning-based model by using a combination of TinyBERT and a convolutional neural network is proposed to detect any vulnerability that may lead to a possible cyber-attack, i.e., ransomware. Simulation results of the proposed study evaluated on a real-life dataset demonstrate that the proposed model effectively detects the ransomware attack with an accuracy of 95.40% as compared to the other state-of-the-art methods.