Research On Anchors Based Localization Algorithm In Wireless Sensor Networks

A flexible, robust, Scalable network called Wireless Sensor Networks (WSN)which established by node self-organization is increasingly coming to scholars’attention with the continuous progress of the practical process. WSN, a core part theInternet of Things technology, have made such a great progress in theoretical researchand practical application (e.g. national defense, environmental monitoring, smart home,vehicle tracking, emergency rescue and et al.). The WSN nodes often random deploy inunattended area with hills, rivers, etc., therefore either open areas or X-type or C-typeparticular network will appear in network boundaries. Thus WSN network structuredesign and related technology research remains challenges to be solved.The accuracy of node location information plays a vital role in WSN applications.Due to the limitation of energy consumption, sensor nodes must get its own positioninformation through localization algorithm, so localization algorithm became one of thehottest WSN researches in recent years. The existing WSN localization algorithms oftendesign for smaller range at indoor environmental and wide range localization algorithmsrarely concern about the special needs of the particular network environment. Inaddition, sensor nodes will not only move its location unexpectedly also face attacksfrom malicious nodes or captured nodes. Therefore, it is particularly important todevelop a well-established reliable node localization algorithm which could apply toparticular network environment.This paper is supported by fund of National Science and Technology Major Project(Project of key technology in electromagnetic spectrum monitoring WSN). On the basisof full investigation and research progress, this paper intensively study three aspects,static node localization, anchor node plausibility verification and dynamic nodepositioning, by using a combination of theory and simulation analysis mode. Mainresearch works are summarized below:First, this paper analyzes the characteristics of existing localization technology, anddescribes a hierarchical distributed network topology of WSN for particular networkenvironment. On the base of indicating paper’s background and significance, wedescribe and analyze the existing localization algorithm research progress from three perspectives which are static node localization, anchor node plausibility verification anddynamic node positioning. After realized these methods characteristics and limitations,line of research is determined by considering characteristics of network topology.Secondly, a static node localization algorithm is studied in this paper. Existedpractical location system often use range-based localization algorithms, however thesemethods not only requires additional hardware devices such as the ultrasonic generator,but also seriously affect the positioning accuracy by various factors which are multipath,uncertainty and nonlinear characteristics, etc. Range-free localization algorithms don’trequire additional hardware support, however, compare to range-based methods in idealsituations, range-free localization algorithms have bigger positioning error, especially inthe particular network environment. To solve above problems, this paper proposes anew method called Range-Free Support Vector Machine Static Node LocalizationAlgorithm (RFSVM) which could apply to the particular network environment. RFSVMcould effectively improve the positioning accuracy by using hops-distancetransformation matrix and support vector machine. And it can be applied to WSN withhigh positioning accuracy, low cost and severe environment, such as geologicalprospecting and other fields.Thirdly, this paper discusses the reliability of the localization process and proposesa new method called Range-Free Anchor Position and Hops Verification Algorithm inWSN Localization (RFPHV). Location information of anchor nodes is the basic ofpositioning algorithm. If malicious or capture anchors send the wrong position or hopsnumber information to other nodes, positioning accuracy of localization algorithm willbe affected greatly. In order to solve this problem, RFPHV is proposed. RFPHV candetermine anchor node plausibility by using network connectivity, hops information anddirectivity of the antenna. The algorithm can not only be associated used in proposedlocalization method, also has good scalability in order to improve the reliability of otherlocalization algorithms.Finally, the dynamic node localization algorithms in the Mobile Sensor NetworksMWSN) is studied in this paper. Dynamic nodes come into WSN in order to meet awide range of its application requirements, and the traditional static localizationalgorithm can’t operate in dynamic network structure. In the existed dynamic nodelocalization algorithms, nodes often have large positioning errors and must move with a particular law of motion. To solve these problems, this paper proposes a FractionalExtended Kalman Filter Dynamic Node Localization Algorithm (FEKF) by combiningKalman filter theoretical model and fractional differential. Compared to other traditionaldynamic node localization algorithm, FEKF can greatly improve positioning accuracy,and apply to many fields in order to satisfy vehicles and personnel dynamic nodeslocalization with characteristics of random movement directions.