Investigation of Deployment Challenges of Sensors and Development of Machine Learning Methods in Wireless Embedded Computing Environment

Abstract

The deployment of the current Internet of Things (IoT) and sensors trend has newlinedramatically risen across all industries from previous ten years. However, much newlineresearch is not addressed on the deployment issues with embedded sensors and newlineassociated wireless networks in terms of contextual dynamics and system newlineperformance. In order to observe the environmental related parameters properly and newlineaccurately from the appropriate embedded sensor system, it is necessary to explore newlinethe possibilities towards the computing systems that uses sensors and their newlinedeployment that perfectly emulate the occurrence of the events in the real newlineenvironment. newlineIn the next generation wireless technologies like 5G/6G and beyond, there is a scope newlinefor intelligent and smart way of information transfer and services to the society, newlinewhich is scalable and reliable. Numerous smart initiatives are being started to address newlinemore intrinsic and extrinsic services in the domains of smart cities, agriculture, health, newlineindustry, automation, etc that requires potential communication capability. Embedded newlinesensors, Internet of Things (IoT), edge computing, and smart sensors are needed in newlineeach of these sectors in order to perceive and tag the phenomena of interest using newlinedigital decision support systems. Recently, research groups have given wireless newlinesensor networks (WSN), which contain several sensors to examine the physical world, newlinea lot of attention and studies devoted to the deployment of sensor nodes and have newlinebeen enhanced by the increasing requirements of WSN in various applications. Here, newlinemost of the node placement models addresses heavily the WSN s optimal coverage, newlineconnectivity and energy efficacy, along with performance evaluation of such systems. newlineThis thesis firstly, proposes a technique for evaluating some of the operational newlineefficiency with respect to energy and their performance with proper placement of newlinesensors. The process entails grouping the sensor nodes according to how far away newlinefrom the phenomena and where it is physically located. The strategies include