Researching On Localization Technology Of Wireless Sensor Networks

There are many problems about localization algorithms of wireless sensor networks currently,such as low positioning accurancy, computational complexity and high energy consumption. Thisthesis introduces the structure of wireless sensor networks, the characteristics of nodes in wirelesssensor networks and the development of wireless sensor networks around the world. Afterdiscussing the node-localization technology, related localization algorithms and wireless signalattenuation models, we combine the RSSI and centroid algorithm and improve it in order toimprove the positioning accuracy.The improved algorithm called weighted centroid localization algorithm based on RSSI. Thealgorithm is divided into tow parts: the first part is the amendment of the RSSI ranging. In order toimprove the accurancy of RSSI ranging, we create a linear correction model to add or avoid themodel according to the principles of the least squares. The second part is centroid localizationalgorithm. It’s the amendment of the weighted factor based on the distance which one is got frompart one. The weighted factor after being improved can response which one is more important forbetter localization. We use MATLAB to set up a simulation platform in order to test whether theamended algorithm can improve the positioning accuracy. Unknown nodes arranged in the area of asquare randomly with some beacon nodes that their location is fixed. We set a series of factors suchas communication radius, node density, path-loss factor and weighted correction factor. Then, wecan come to a conclusion from the simulation picture that the amended algorithm can get a betterresult by changing one of the paramenters.Applying some mathematical models, such as the least squares model and weighted facoter, tonode-localization in wireless sensor networks can reduce error and improve accuracy.