Fuzzy‐Driven Cluster Head Selection and Deep Learning Prediction on the Basis of Hybrid Optimization Algorithm for Multiobjective Routing in WSN‐IoT
R. Ramya, Thomas Brindha · International Journal of Communication Systems · 2025
ABSTRACT In the Internet of things–aided wireless sensor networks, the deployment of several nodes on a large scale presents distinctive complexities different from conventional WSNs, which leads to varied issues and challenges. Generally, the sensor nodes operate on restricted energy resources, and they play a vital role in the networks. Designing a protocol that is both resilient and energy efficient to enhance the network's prolonged existence poses an important issue. Thus, a novel metaheuristic optimization algorithm is introduced in this paper, which is aimed at addressing cluster head selection and routing in IoT‐based WSNs. Generally, selecting cluster heads is considered essential for enhancing the performance of clustering routing protocols. To address this, a novel approach for optimization called the Artificial Hummingbird Cheetah Optimizer Algorithm is presented, which is the combination of both the Artificial Hummingbird and Cheetah Optimizer Algorithm. Additionally, multiobjective prediction is conducted by employing a convolutional neural network model, which is trained with a cat‐and‐mouse optimizer model. Then, CHS is performed using a fuzzy system, by considering predicted energy levels, lifetime of the sensor nodes, trust, and quality of service. Routing is next executed by employing the AHbCOA model. Finally, an experimentation analysis is performed by evaluating the performance of the presented approach with several optimization models. The performance of these approaches is evaluated by employing various metrics such as energy consumption, throughput, QoS, trust, and LLT. The overall analysis reveals that the proposed model outperforms others, achieving values of 0.933 (J) for energy, 0.592 (s) for LLT, 0.938 for trust, 0.791 (Mbps) for QoS, and 0.934 (kbps) for throughput.