Spiking Residual ShuffleNet-Based Intrusion Detection in IoT Environment

The Internet of Things (IoT) system has been developed to create a smart environment. Privacy and security are critical issues in IoT systems. Security vulnerabilities in IoT-enabled models generate threats that impact various applications. Therefore, intrusion detection systems (IDS) are essential for IoT systems to prevent attacks. Due to the limited storage and computational abilities of IoT devices, traditional IDS is not suitable for IoT systems. In this context, a new model called Spiking Residual ShuffleNet (SR-ShuffleNet) is introduced for intrusion detection (ID) in an IoT environment. The Multi-objective Fractional Artificial Bee Colony (MFABC) algorithm is used for routing. MFABC combines fractional calculus with artificial bee colonies. At the base station (BS), data normalization, data augmentation (DA), feature selection (FS), and ID processes are carried out. The input data is normalized using Z-score normalization (ZN). Features are selected based on Motyka similarity and Topsoe similarity. After feature selection, the bootstrapping technique is used to augment the features in the DA phase. Finally, ID is performed using SR-ShuffleNet, which integrates the Spiking deep residual network (S-ResNet) with ShuffleNet. After detecting an intrusion, an attack is mitigated to eliminate the malicious node. The SR-ShuffleNet model achieves impressive accuracy and efficiency in identifying and mitigating attacks in IoT systems.