Coverage Performance Analysis of Wireless Networks in Presence of Sensor Failure

Abstract

newline A dramatic advancement in Wireless Sensor Networks (WSNs) application domain, including newlineborder region monitoring, surveillance, reconnaissance, precision agriculture, industrial newlinemonitoring, etc have been seen in last decade. The performance of these networks for various newlineapplications depends on many factors such as characteristics of signal propagation environment, newlinethe shape of and size of the deployment region, Boundary Effects (BEs), SF, sensing and newlinetransmission capability of the Sensor Nodes (SNs). Since, these networks are deployed in hard-toand#65534;reach areas such as forests on fire, earthquake-prone areas, battle fields, and many others. Also, the newlinedeployment of these networks involves a huge capital investment. Therefore, it becomes necessary newlineto formulate analytical models that can estimate these network coverage prior to their installation newlinein the region of interest (RoI). newlineIn this study, we present analytical models to compute the coverage of finite WSNs deployed newlinein finite regions considering BEs, SF, and the characteristics of the propagation environment. We newlineobtained analytical solutions for and#120581;-coverage probability (and#120581;-CP) of WSN spread in a given newlinerectangular/circular shaped finite region. Using the proposed work, we can estimate the impact of newlinevarious network and environmental factors such as SN s count, SR, SF, and the Shadowing Effects newline(SEs) on the and#120581;-coverage metric (and#120581;-CM). The proposed analytical models are validated through newlineextensive simulations, and the results obtained through simulations conform the analytically newlineobtained results, thus, substantiating the presented models. However, it has been observed that the newlinevalidation of analytical solutions through simulations is a monotonous and laborious task taking newlineseveral hours for a single outcome. In addition, the validation through the practical implementation newlineneeds a huge amount of financial support, which is a major issue. newlineTo overcome these problems, this work proposed a ML technique based on a fully connected newlinefeed