Study On Localization Algorithm Of Wireless Sensor Network

Wireless sensor network(WSN) is a novel network technology about acquiring andprocessing information. It can be widely used in a multitude of diverse applications, suchas industry, agriculture, environment monitoring, military field and so on. Locationinformation has played an extremely important role in these areas. Although the way to setup manually or install GPS can obtain location information accurately, it will waste humanresources and a vast amount of time. A better way to get position information is to use thelocalization algorithm which is appropriate to the characteristics of WSN. This paperfocuses on the research of wireless sensor network localization algorithm.Firstly, in this paper, the research background, significance and status for WSN areoverviewed based on a number of related literatures. And the framework, characteristics ofWSN and traditional classical localization algorithm that range-free are introduced.Secondly, the questions are happened in the process of APIT test, such as boundaryeffect, the accuracy that determining the unknown node belongs to which area is notenough, low fault tolerance of the grid scan algorithm. Region divided-based localizationalgorithm is proposed. Compared with the traditional APIT algorithm, the proposedalgorithm improves three places. Firstly, the improved method adds a counter for eachunknown node. Secondly, The triangle of PIT test is divided into several sub-areas. Thirdly,the position of the unknown node is estimated by triangle-center scan algorithm. Finallythe purpose is to get more accuracy.Thirdly, another improved algorithm is proposed to make up for the shortage ofcentorid algorithm, which is in condition of uneven distribution of anchor nodes. Rotationangle test method is proposed in the improved algorithm. This method optimizes theunknown node localization effect for anchors and reduces the position error brought byuneven distribution of anchor nodes.Lastly, these proposed algorithms are tested by simulations via Matlab and arecompared with traditional classic localization algorithms according to localization error.