Empowered MapReduce and Deep Trust Management for Mitigation of Void and Energy Hole in WSN-IoT

Abstract

AbstractThe most prevalent challenges in wireless sensor networks (WSNs) are occurrence of void holes and energy holes, which shorten the lifetime of sensor nodes and cause rapid energy depletion. Most of the existing literatures are not suited for energy depletion due to indeterminacy in selection of cluster heads and uneven distribution of sensor nodes in WSN. The optimized clustering in presence of noisy and border lying sensor nodes is the main disadvantage while using the standard energy hole alleviation algorithms. This paper improves the process of alleviating void hole occurrence in WSN by devising a wedge-based partitioning to distribute the sensor nodes evenly. The standard Fuzzy C-Means algorithm’s efficiency is low when dealing a huge volume of traffic with limited memory. Another difficulty is how to optimize Fuzzy C-Means clustering is a major challenge and it is achieved by integrating MapReduce framework to improve clustering reliably. The MapReduce based Fuzzy C-Means clustering is developed in this research work to determine the optimized cluster head in presence of indeterminacy due to the dynamic nature of the environment and mobility of sensor nodes. Hence, it can perform parallel clustering of sensor nodes to improve the speed of the algorithm. Additionally, security is also considered as an important factor to accomplish better energy consumption in WSN. This proposed work introduced a deep neural network-based trust management scheme to overcome the issue of data packet loss and time delay. The simulation results proved that the proposed Empowered MapReduce and Deep Trust Management (EMR-DTM) with its ability of tristate mechanism greatly overcomes the problem of energy depletion and secure data transfer more precisely compared to other standard state-of-the-art clustering algorithms in WSN-IoT.KeywordsWireless sensor networkInternet of ThingsWedge modelFuzzy C-Means clusteringMapReduceDeep neural networkTrust managementEnergy hole and void hole