RF signal modeling and deployment strategy targeting outdoor RSS - based localization and tracking applications in wireless sensor networks

Abstract

The localization of the sensor nodes is a fundamental issue in the area of wirelesssensor networks (WSNs). An attractive way for estimating the location of mobile orstatic wireless objects is by using the received signal strength (RSS) attenuation withthe distance, which does not require any additional hardware. This is possible due tothe fact that in most sensor nodes radios the received signal strength indicator (RSSI) isa standard feature and can be obtained automatically by the received messages. On theother hand the RSS is known for being noisy, unstable, variable and difficult to use inpractice. For achieving a better understanding of the nature of these difficulties andlimitations, and for identifying the range of applicability of the RSS in localization andtracking scenarios, a thorough study about the RSS and its dependence on the variousfactors and environmental conditions is essential.The present doctoral dissertation investigates the feasibility of sensor nodelocalization and target tracking with the resources of the WSN technology, when usingonly the RSS of the exchanged messages. Moreover, it offers experimental support tothe hypothesis that proper modeling of the RSS behavior and appropriate selection ofthe topology parameters are essential for the applicability of WSN in real worldconditions. In brief, the present doctoral dissertation concerns with: (i) identifying themain factors that influence the accuracy, the variability and the reliability of theobtained RSS, (ii) modeling the RF signal propagation in the context of WSNs, and (iii)defining the basic deployment constraints and evaluation of the topology parametersthat can guarantee successful localization and tracking.For assessing the practical value of various RF-models, experiments using TmoteSky and TelosB sensor nodes in real-field outdoor environment were carried out. Theimpact of a number of factors, such as the operating frequency of the radio, thetransmitter–receiver distance, the variation of transceivers hardware due tomanufacturing tolerances, the antenna orientation, and the environmental conditions,on the RSS was investigated. The influence of the various factors that affect the RFsignal propagation and some constraints imposed by the WSN nature was accountedin order to design practical models, suitable for outdoor unobstructed and outdoortree-obstructed environments.A pre-deployment simulation framework has been introduced and in its context aRF signal propagation-based connectivity strategy (RFCS) has been developed to fulfillvthree deployment provisions: (i) discovering the most appropriate height from theground and distances for the sensor nodes, (ii) reducing the transmission power, and(iii) minimizing the interference from non-neighbor nodes. The RFCS uses a RF signalpropagation model to predict the RSS in order to identify the most appropriatecommunication-based deployment parameters, i.e. T-R distance, height from theground and transmission power.The localization and tracking considerations, by means of localization and trackingtechniques, topology parameters and factors influencing the localization and trackingaccuracy, are combined in illustrative simulation examples to evaluate theirsignificance concerning the performance of the localization and tracking task.Furthermore, the propagation model and the topology parameters being identifiedwere validated in real outdoor sensor node localization and target tracking tests