A Pragmatic Approach of Location and Tracking Algorithms in Wireless Sensor Networks

Student thesis: Doctoral thesis

Abstract

The last advances in wireless communications and electronics have motivated the appearance of Wireless Sensor Networks. These networks are formed by a new kind of low-power and low-cost sensors able to operate across short ranges. Their simplicity and autonomy have motivated the development of many final applications in a large variety of fields. Nevertheless, sensor nodes are equipped with limited data processing and communication capabilities. Hence, several design challenges appear when an application has to be developed. These restrictions justify the design of highly distributed and energy-efficient applications. Localization and tracking algorithms are one of those emerging applications that have become an interesting field to the researchers. The information routing is often supported by their localization. Besides, the location knowledge gives to the data sensed a geographic sense. Instead of using the existing global localization methods, such as GPS, that are more complex and costly, recent advances have demonstrated the viability of local methods. In this PhD dissertation, we have focused our study of the localization and tracking algorithms for WSN on the RSS-based distributed approaches. One of the major issues is to obtain the simplest possible method, and RSS range measurements have become the simplest existing measurements. Besides, we have also presented methods that are able to optimize the trade-of between accuracy versus energy-efficiency. First, RSS-based cooperative localization algorithms in static indoor networks are considered. The use of RSS measurements requires the knowledge of a propagation model in order to obtain inter-node distance estimates. We introduce an on-line path loss estimation method that obtains the model by means of RSS measurements. Hence, we avoid the need of an a priori estimation of the propagation model. Moreover, the cooperative approaches used increase the number of nodes that cooperate with a non-located node in the location estimation procedure. Two major issues have to be taken into account when a large number of nodes are used. On the one hand, the larger the number of cooperating nodes, the larger the number of messages exchanged, and, hence, the higher the energy consumption. On the other hand, the probability of using further nodes is increased, hence, the higher the distance, the higher the error distance estimates, when RSS measurements are used. These features have motivated us to propose three different node selection criteria in order to reduce the energy consumption maintaining the accuracy. Finally, we have considered the mobility of the non-located nodes inside a fixed network. The interest is to locate and track a node moving across a WSN. We have considered two different scenarios: an outdoor one, in which the velocity is medium-high, and, an indoor one, where the velocity is lower. In both cases, we have still used an RSS-based cooperative algorithm. Besides, we introduce the Kalman Filter and its derivatives, because, they have become a common approach used for tracking purposes. In both scenarios, the mobility of the node causes a high variability of the RSS measurements. These errors reduce the accuracy. In that sense, we propose a window-based RSS correction method in order to counteract these negative effects.
Date of Award1 Oct 2012
Original languageEnglish
SupervisorJose Lopez Vicario (Director) & Gonzalo Seco Granados (Director)

Cite this

'