Study Of Three-dimensional Localization Method Based On RSSI In Wireless Sensor Network

With the development of wireless sensor networks (WSN) technology，it’s widelyapplied to many fields such as national defense, environmental monitoring, trafficcontrol and medical treatment. In various applications of wireless sensor networks,the location information is critical for monitoring activities. But in most cases, sensornodes are randomly deployed in the monitored area, and their positions areunpredictable. Such as environmental monitoring in oceans and lakes, each node getswater temperature, salinity, pressure and flow in a different depth. Because of severalof factors, such as the hardware cost or environment, positioning methods relied onglobal positioning system (GPS) can not fully meet the needs of people. As thedemand for location-based applications improved step by step, two-dimensionalpositioning could not meet the requirement of practical applications. Sothree-dimensional positioning is needed. The research of this paper comes from theProvincial Natural Science Foundation. The study of this paper is on wireless sensornetwork nodes own spatial orientation based on the Received Signal StrengthIndicator (RSSI).Three-dimensional localization algorithm for wireless sensor network ispresented comprehensively in this paper. A novel refinement algorithm based on LQIconfidence for three-dimensional localization (3D-RABLC) is proposed in study ofthe problem of gross error appeared from RSSI-based ranging method. Firstly,according to cluster structure model, a positioning region is divided and built for thecluster positioning unit area in lots of nodes deployed randomly. Secondly, the pathloss exponent is calculated by using anchor nodes, and the relationship of the distancebetween nodes could be attained through the RSSI signal model transformation.Thirdly, RSSI value degree of confidence is divided from the research of Link QualityIndicator (LQI). The lower confidence RSSI is modified through the higher to buildthree-dimensional hops modeling or making use of compensating method. In the last,Multidimensional Scaling (MDS) centralized algorithm is used to attainthree-dimensional relative coordinate among nodes, then the three-dimensionalcoordinate relationship could be transformed through three-dimensional vector spacecoordinate in the global.In this paper, a position test bed based on LQI of the3D-RABLC algorithm isconducted completely indoor and outdoor. The results of this experiment show that the algorithm reduces the RSSI rang error and improves the accuracy ofthree-dimensional localization significantly.