Font Size: a A A

Study On Node Localization Algorithms In Wireless Sensor Network

Posted on:2012-05-28Degree:DoctorType:Dissertation
Country:ChinaCandidate:Y LiuFull Text:PDF
GTID:1118330335952048Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
Wireless sensor networks are composed of dozens or even thousands of sensor nodes which are deployed in a wide and complex area, some of which men can not reach that working together to monitor a region to obtain data about the environment. Sensor nodes may be deployed by means of random scatter done by aircraft or people. Generally, measurement data is collected by sensors and delivered to center node or processor to analyses. However, the collected data are meaningless without corresponding position information. When sensor networks are established, several nodes are specially arranged, which are called anchor nodes and their location information should be clear. Anchor nodes can do self-positioning by GPS or other means, but the cost of using GPS is higher than that of common sensor node. Thus, designers can not make each sensor node possess the self-positioning ability. So the various localization technique of wireless sensor network emerges as the times require. The principle goal of WSN is that How to realize the high-accuracy to meet the application.At present, the node localization has become a very hot issue in the research of wireless sensor networks. This paper deals with both range-based and range-free localization algorithm in wireless sensor network, and analyses and summarizes the advantages and shortcomings of these two types of localization algorithms, as well as the main factors affecting the localization accuracy. Among them, an anchor node is a critical factor in localization algorithm, as every localization algorithm needs anchor nodes to realize localization. The effect of anchor node upon the location accuracy embodies in two aspects, specifically the number of anchor nodes and position relationship between them. Therefore, the improvement the location accuracy by some novel equitant to anchor nodes derived on the basis of the existing anchor node, and the analyzed position relationship between the anchors nodes are the main difficulties of localization algorithm. Most of currently research focus localization algorithm, but on estimation of error through verifying the position of nodes researches are few. Hence it is critical for application of WSN to find an evaluation system reflecting the advantages and shortcomings of localization. The TDOA algorithm is a range-based localization technology, and the main cause of error for this algorithm comes from the time difference of measuring, which is determined by hardware conditions and environment factors, and is impossible to eliminate. Hence it can only be optimized from the solution of TDOA localization algorithm. Aiming at the existing problems in current methods, a series of new algorithms and corresponding solutions for location algorithm is also proposed. The innovation of this paper lies in the following four aspects:1. The concept of the Voronoi diagram from computational geometry was applied to range-free localization algorithm of wireless sensor network, and a node localization algorithm based on Voronoi diagram has been proposed. Without the increasing of location cost and not use of auxiliary hardware, add several anchor nodes information (virtual anchor node) equivalent for location space are studied, so as to increase the number of anchor nodes communicating with the sensor nodes and further increase the node information determining the position itself. Meanwhile a sequence level method to construct anchor nodes, virtual anchor nodes and unknown nodes is presented and unknown nodes finally realize localization. The plan effectively improves the location accuracy, which don't follow the traditional algorithm of adding the number of anchor nodes at the cost of increasing the network cost and energy consumption.2. It was the first time that the gradient concept of geometry was applied to the wireless sensor network localization algorithm. The method for analyzing the position relationship between anchor nodes has been investigated and given the location of geometric algorithm. In two-dimensional plane, locating a target node requires three anchor nodes, which has a significant selected position. If the selected anchor node doesn't fit, this can not locate the position of unknown node. Therefore, before localization, selecting the optimum anchor node combination in accordance with the proposed scheme, and then locating the node can improve the accuracy of localization. This method provides reliable basis for the selection of anchor node in localization algorithm to improve location accuracy and location coverage.3. The reliability of localization algorithm was tested by using probability detection model. The size of location errors can significantly determinate the application of localization algorithm. If the position error is too large, the false location information would be transmitted to the monitoring center. This kind of misleading could result in enormous losses for wireless sensor networks application. This method provides a strong technical support for wireless sensor network application. This model will analyse the DV-Hop localization system. Under the condition that the real position of the node is unknown, the node position error can be judged according to the probability model and the optimal and worst of node localization can be found out, all of which complete the evaluation system of WSN in application.4. The TDOA localization algorithm of range-based was discussed performance. The accuracy of TDOA location is determined by the size of errors in time differences. If the error of measuring time value is too large, according to TDOA equations can not locate the position of unknown node. Therefore the combination of the classical Newton's iteration with TDOA was made and dual nonlinear equations were constructed. In addition, the TDOA Newton iteration of hyperbolic equations was derived theoretically, which can solve the problem of hyperbolic equations due to the measurement errors resulted from time differences. It established the theory foundation for the development of the location technology of wireless sensor network.In summary, by researching on the number of anchor nodes, the distribution location of anchor nodes, the location error analysis and range-based optimization algorithm, this paper solves the key problems existing in the localization of WSN from several angles, and played a promoting role application and development of wireless sensor network.
Keywords/Search Tags:Wireless sensor network, localization, voronoi diagram, gradient, detection probability, Newton iteration
PDF Full Text Request
Related items