DEVELOPMENT OF AN EFFICIENT LOCALIZATION SCHEME INWIRELESS SENSOR NETWORKS USING MACHINE LEARNING

Abstract

WSN applications have paved the way for creative research in the telecommunications industry. One of the crucial WSN challenges to investigate with the use of localization techniques is the localization of nodes. Localization strategies or procedures refer to the processes used to estimate the position coordinates of sensor nodes. The geographic positions of sensor nodes in a WSN are referred to as location. Drones, cell phones, etc. are a few examples. Both absolute and relative locations are possible. The node's precise coordinate is measured by the absolute location. In situations like these, like Battlefields, localization inaccuracy is not acceptable. Locations are measured relative to another node, such as a wildlife application, a fire detection system, etc. The necessity for localization is brought on by the maintenance of sensors and the redeployment of nodes, respectively. There are two different kinds of nodes in a sensor network: anchor nodes, which are aware of their location, and unknown sensor nodes, which are unaware of it. The localization algorithms have a number of problems, including the need for a small number of anchors, latency, efficient energy use, precise location estimation etc. Machine learning in WSN generates models to automatically and accurately analyse complicated data as well as increases computing process efficiency, cost-effectiveness, and reliability. The performance of the network can always be improved by using machine learning techniques. It limits the requirement for human interaction and the necessity to constantly programme after errors. It makes it simple to access and interpret a large volume of data that has been periodically accumulated over time by sensor nodes.