A DEEP LEARNING APPROACH FOR TRUST-AWARE AND FAULT-TOLERANT SECURE ROUTING IN IOT USING MC-GRU AND GREYSTAR GOOSE OPTIMIZATION

Abstract

Internet of Things (IoT) is a system of physical devices that are interconnected to collect and exchange data worldwide using the Internet. Various smart devices are embedded in IoT with software, sensors, and other technologies, like smart lights and thermostats, to complex industrial tools. However, the dynamic and massive IoT environment encounters significant challenges related to trust, reliability, and security. The existing secure routing protocols fail to effectively ensure data privacy, while maintaining energy efficiency. This paper introduces GreyStar Goose Optimization Algorithm (Grey-SGOA) for trust-aware and fault-tolerant secure routing in IoT. Here, the IoT network is simulated initially, and then, clustering is performed by K-medoids clustering algorithm. Then, the cluster head is selected using Grey-SGOA by concerning different fitness parameters and trust factors. After that, the data aggregation is performed using Multi-Context Gated Recurrent Unit (MC-GRU), and the encryption of aggregated data is executed using Fully Homomorphic Encryption (FHE). Secure routing is performed in the IoT network using Grey-SGOA. Further, the efficacy of Grey-SGOA is investigated, where the Grey-SGOA attained average residual energy, trust, delay, Packet Delivery Rate (PDR), throughput, normalized variance, and conditional privacy of 0.274 J, 81.739, 79.131 msec, 89.651%, 145.045 Kbps, 0.008, and 82.428%.