Enhanced Network Lifetime and Secure Data Transmission in IoT: A Reinforcement Based Approach

Internet of Things is a developing technology in the modern period with uses in wide area monitoring, healthcare, smart cities, and other areas.Wireless sensor networks (WSN) form the core of these IoT.The sensors within the WSN encounter numerous obstacles, including limited battery life, Security and Privacy and Synchronization etc.This shortens the lifetime of the network, thus energy must be used wisely.In this study, the Reinforcement Learning algorithm and K-Means are used to build clusters and to elect cluster heads.Additionally, a mobile sink is proposed to collect data from each cluster head (CH), saving energy on data transmission from nodes to base station.The proposed novel clustering and cluster head election algorithms increases energy efficiency by 75% and the time complexity is reduced to O(n/2) using reinforcement algorithm.By considering the shortest edges of obstacles, energy consumption during the mobile sink's routing is reduced, ensuring secure data transmission.The results illustrate a comparison of the time complexity between the initial cluster head election, conducted using K-Means, and the subsequent cluster head election is performed using Q-Learning.These algorithms are compared with K-Means and Fuzzy C-Means.Our proposed approach demonstrates superior performance compared to the other methods.The outcomes of the proposed work also reveal that in comparison to Clustered Routing algorithm based on forwarding mechanism optimization (CRFMO), Low Energy Adaptive Clustering Hierarchy with Improved Adaptive Cluster Adjustment (LEACH-IASA) and Improved LEACH, enhances network lifetime by increasing the number of surviving nodes and network coverage.The proposed scheme also outperforms in terms of latency compared to Software Defined Networking with Reinforcement Learning (SDN-RL) and Energy Efficient Rendezvous Ponts Selection using Deep Policy Dynamic Programming (EERPS-DPDP).