QoS-Aware Resilient Routing Protocols Leveraging Cosine Similarity-Centric Convolutional Neural Network for WSN-Assisted IoT Using Clustering Techniques

The Wireless Sensor Networks integrated with Internet of Things (WSN-IoT) serve as the backbones of such applications as smart cities, health monitoring, and environmental monitoring requiring high efficiency and secure communication for Quality of Service (QoS). However, an important challenge has been in routing in dynamic WSN-IoT systems ensuring energy-efficient resilient and QoS-aware operations. This research addresses the issues of hotspot formation, energy depletion, and secure data transmission in clustered WSN-IoT networks. The requirement for robust scalable routing protocols able to maintain QoS requirements even in dynamic environments drives the motivation to develop new protocols. The proposed “QoS-Aware Resilient Routing Protocols leveraging Cosine Similarity-Centric Convolutional Neural Network for WSN-assisted IoT Using Clustering Techniques” (Ski-CSC2-A2O), leverages a Bi-Concentric Hexagonal network structure with mobile sink assistance for energy-efficient data collection. The Skill Optimization Algorithm optimizes clustering which results in balanced power consumption while identifying optimal Cluster Heads. Secure and QoS-aware routing is achieved through a Cosine SimilarityCentric Convolutional Neural Network, while network parameters are fine-tuned using Aphid Ant Optimization. Simulation results demonstrate the effectiveness of the proposed protocol, achieving a Packet Delivery of 99.3%, throughput of 99.6%, and an extended network lifetime surpassing 99.4% of baseline approaches, even with large packet sizes and increased transmission rounds. The Ski-CSC2-A2O protocol provides WSN-IoT systems with highly efficient secure solutions through its capabilities to ensure energy efficiency robust performance QoS while functioning ideally in real-time IoT applications.