Analysis of machine learning based LEACH robust routing in the Edge Computing systems

Abstract

Wireless sensor networks (WSN) are used to detect real-time changes in the deployed environment. This dynamic behaviour is either triggered by the deployed environment or by the user from outside. Because of their ability to monitor complex scenarios that change rapidly over time, wireless sensor networks are critical components of most advanced computing systems. These complex activities are influenced by different methods or even by the designers of their networks. Machine learning encourages many real solutions that optimise resource use and increase the network's lifespan in sensor networks. LEACH routing protocol has many limitations due to sudden energy utilisation & cluster head nodes due to direct communication with the base station node. This fast node energy leak creates several black hole structures in the networks, resulting in data redundancy, data packets transmission, node upgrade costs, and end-to-end delay for WSN. The proposed model with LEACH protocol functionality has improved network performance, network (WSN) efficiency, and solving data redundancy issues. By using an independent Recurrent Neural Network (IRNN)-based data fusion algorithm, namely, DFAIRNN. The simulation and comparative results indicate that the mean method & minimum distance method used in the LEACH-DFAIRNN protocol can effectively resolve data redundancy issues caused by the adjacent sensor nodes by flooding data simultaneously to a single node.