A Reliable Hybrid Framework for Anomaly Detection in Secure and Robust Wireless Sensor Networks

The integration of the FireTG-Net model enables Firefly Swarm Optimization (FSO) to work with the Temporal-Gated Recurrent Unit-Network (Temporal-GRU-Net) for detecting anomalies in Wireless Sensor Networks (WSNs) resulting in 98.13% detection accuracy. The model employs FSO-optimized trust evaluation for local and global assessments that enables it to automatically adjust its detection methods to changing conditions and improve the efficiency of routing decisions. The model identifies immediate and changing malicious behaviors of blackhole, sinkhole, and jamming attacks through its 1D convolutional layers and advanced Gated Recurrent Units (GRUs) feature extraction capabilities. The FireTG-Net model exhibits superior evaluation scores compared to the Decision Tree (96%), Fuzzy Model (81%) and Trust-aware Routing Protocol (TRP) (96.791%) while facing disruptions better and showing restricted latency effects. FireTG-Net demonstrates excellent capabilities for enhancing the security and reliability of WSNs through its effective performance regarding high packet delivery ratio, enhanced energy efficiency, and reduced false positive rates.