Dynamic routing protocol using fuzzy based deep reinforcement learning for internet of things enabled wireless sensor networks

Abstract

The constant movement of sensor nodes adds complexity to the task of data newlinerouting across several access points. So, it brings about node selection mistakes, an newlineinability to prolong the lives of particular nodes, reaction time delays, packet loss, and newlinean increase in computational complexity. The current study presents a deep newlinereinforcement learning (DRL)-based intelligent data routing technique to enhance the newlineIoT-enabled WSNs performance by taking into account parameters including temporal newlinecomplexity, maximum data sum rate, and message overhead. Using a double cluster newlinepairing strategy, the initial instantaneous data load is split across pairs, with each pair newlineconsisting of two sensor nodes-one strong and the other weak. newlineThe approach presented by this study offers the benefits of state-of-the-art newlinerouting techniques such as prolonging the lives of nodes, lowering less power newlineexpenditure, and being implementable on any network platform, having mobile and newlinenon-mobile nodes. The energy consumption of Wireless Sensor Networks (WSNs) is newlinebeing constrained by their batteries. In this thesis, we suggest a strategy to reduce the newlinepower needed to route the network s various components. OEERP stands for newlineopportunistic energy-efficient routing protocol and is a concept that may enhance newlinenetwork performance, target location detection precision, network longevity, and newlineenergy efficiency. newline