DeepLG SecNet: utilizing deep LSTM and GRU with secure network for enhanced intrusion detection in IoT environments

• Published in: Cluster computing Vol. 27; no. 4; pp. 5459 - 5471
• Main Authors: Nanjappan, Manikandan, Pradeep, K., Natesan, Gobalakrishnan, Samydurai, A., Premalatha, G.
• Format: Journal Article
• Language: English
• Published: New York Springer US 01-07-2024; Springer Nature B.V
• Subjects: Access control; Algorithms; Computer Communication Networks; Computer Science; Context; Cybersecurity; Data integrity; Datasets; Deep learning; Effectiveness; Information sharing; Internet of Things; Intrusion detection systems; Neural networks; Operating Systems; Optimization techniques; Performance evaluation; Processor Architectures; Smart buildings; Smart cities; Trends; Internet of things; Intrusion; Gated recurrent units; Long short-term memory; Anomaly; Cybersecurity; Denial of service
• ISSN: 1386-7857; 1573-7543
• DOI: 10.1007/s10586-023-04223-3

The rapid proliferation of the Internet of Things (IoT) has led to a significant surge in interconnected devices across diverse domains, ranging from smart homes and healthcare systems to industrial automation and smart cities. However, this exponential growth has exposed IoT devices to a plethora of cyber threats, including illegal access, data breaches, and malicious attacks, primarily due to their inherent limitations in terms of network capabilities, computational power, and memory. To combat these security challenges and ensure the safety of IoT ecosystems, the development of effective intrusion detection systems has become imperative. Such systems play a crucial role in detecting and preventing unauthorized activities within IoT networks. In this context, this article presents a pioneering approach called DeepLG SecNet, which leverages a combination of deep learning techniques, including Long Short-Term Memory (LSTM), gated Secure Network (SecNet), and Crossover Chaos Game Optimization (CCGO), to fortify IoT devices against unauthorized access and potential threats. To validate the efficacy of the proposed DeepLG SecNet method, various samples were collected from the BoT-IoT dataset and the NSL-KDD dataset. Performance evaluation was conducted using essential metrics to assess the model's detection capabilities in an IoT intrusion context. The experimental analysis yielded promising results, highlighting the effectiveness of the DeepLG SecNet method in intrusion detection for IoT environments. Specifically, DeepLG SecNet outperformed existing methods, demonstrating higher accuracy, precision, recall, and F1 score in safeguarding IoT systems from potential security breaches.