Research Of Node Self Localization In Wireless Sensor Network

As a new technology of gathering and handling information, wireless sensor network has widespread application. It has caused the attention of each domain in many countries all over the world. Due to the fact that positional information which is determined by the position of sensor node is necessary in massive applications of wireless sensor work, the research of sensor node self-localization technology has both theoretical and practical significance.Based on the deep research in the theory of wireless sensor network and sensor node self-localization technology, this paper combines node self-localization technology and the subject of the library's research--rescuing of mine disaster. According to the application scene after the disaster occurs, wireless sensor network is laid out on ground to detect and locate the sound signal transmitting in the stratum. This paper analyzes the actual environment and performance requirements; chooses the algorithm united by range finding method based on signal strength and trilateral locating as the object of research.All the information of neighbor anchor nodes is used to calculate the unknown nodes'coordinates in traditional algorithm locating algorithm based on RSSI. This paper puts forward the idea of screening to the neighbor anchor node first, abandoning the anchor nodes far away and only the information of nearest four anchor nodes is used to calculate a number of coordinates to carry on the coordinate revision. Line error model is proposed to determine the error between the real distance and the calculated length between the unknown node to it's anchor node. Assuming the line error model of nodes is the same if the nodes are located by the same anchor nodes, the error model is determined by calculating the error between the real distance and the calculated length between the anchor node and another nearby anchor node. The unknown nodes'coordinates are modified with the error. For the situation that some nodes fail to be located due to lack of enough neighbor anchor nodes, this paper brings forward the idea that the non-anchor nodes which have already been located successfully could be upgraded as anchor nodes used to locate other unknown nodes. Considering the possible iterative error which may occur, a corresponding weight value is set to the nodes which have already been located successfully each time to make sure that the nodes which are located successfully first is the first choice with the same distance to an unknown node as other anchor nodes when unknown nodes will be located. So the accumulated error is reduced.MATLAB is chosen to establish the ideal two-dimensional plane model to simulate the performance of the algorithms. Results indicated that the adjusting algorithm is better than traditional algorithm in accuracy of pointing. Besides, the algorithm of non-anchors'upgrade highly improves the coverage rate of location.In order to be much closer to actual application scene, this paper applies the algorithms to three-dimensional space, adds the height coordinates of the nodes and finally puts forward the cooresponding 3D localization algorithms. Likewise, three-dimensional space model is established by MATLAB in random to verify the validity of the algorithm in three-dimensional space.