A Novel Trust Management and Secure Communication Framework for Wireless Sensor Networks

In Wireless Sensor Networks (WSNs), ensuring secure and reliable communication amidst various cyber threats is a pivotal challenge. Existing security methods often struggle with high computational demands and do not adequately address the unique characteristics of WSNs, such as the limited energy resources and their susceptibility to specific types of attacks like blackhole and Sybil attacks. The proposed Lightweight MG-Net Model addresses security and performance challenges in WSNs by integrating a Trust Model, Anomaly Detection, and Secure Communication protocols into a novel hybrid deep learning framework. This framework combines MobileNet, which utilizes depthwise separable convolutions for efficient spatial feature extraction, with Gated Recurrent Units (GRUs) to capture temporal dependencies, enabling precise real-time anomaly detection with reduced computational demands. Trust management leverages a modified EigenTrust algorithm, dynamically updating trust scores based on node interactions to optimize reliability across network operations. The anomaly detection component was rigorously trained using a labeled dataset that includes various attack scenarios such as blackhole attacks, where detection accuracy exceeds 97.5%, and Sybil attacks, highlighting its robustness against sophisticated threats. Secure communications are upheld by the Datagram Transport Layer Security (DTLS) protocol, ensuring data integrity and confidentiality with an encryption success rate of 97%. Operational performance metrics are evaluated through simulations, showcasing the system’s efficiency with a detection latency under 2 s and energy consumption that is 30% lower than traditional security frameworks. Overall, the Lightweight MG-Net Model enhances WSN security without compromising on efficiency, demonstrating significant advancements in trust management, anomaly detection accuracy, and secure, low-latency communications.