Research And Improvement On Node Localization Algorithm In Wireless Sensor Network

In recent years, the industry and academia have been great concerned with wireless sensor network because of its wide application prospects in the field of national defense, health care, environmental monitoring, industrial manufacturing and transportation management. Wireless sensor network has integrated a lot of advanced technologies such as embedded technology, microelectronic technology, wireless communication technology, modern network and distributed information processing. It is the production of traditional disciplines and information technology and is a brand new direction in the field of information science. It can monitor and sense and gather information from objects (such as temperature, sound, humidity, pressure, etc.) in a variety of environments. Then it will process the information collected and send them to the users or researchers wirelessly. In the meantime, the precondition of using the information is the accurate node-location. If wireless sensor network can not provide node-location, the information collected will be meaningless. Therefore, localization technology is one of the key technologies in wireless sensor network.Recently, researches for node localization can divide into two models. One is based on range and the other is without range. Localization algorithms based on range have lower localization error and are mainly affected by range-error. But requirement of these algorithms on node’s hardware is higher and calculation is more complicated. So these algorithms are not applicable for low-power, low-cost system. Localization algorithms without range however have higher localization error and have lower requirement on node’s hardware and more simple calculation.This paper has analyzed the classic localization algorithms in wireless sensor network and proposed two improvement algorithms based on centroid localization algorithm. Algorithm one is based on range-free model and two is based on range-based model. Algorithm one named as linear-regression-based weighted centroid localization algorithm has improved centroid localization algorithm through four steps which are weight, centralization, decentration and amendment. Using distribution rule showed by localization error of improved weighted centroid localization algorithm and linear regression model, algorithm one tries to reduce node’s localization error and increases localization ratio. It has been appropriate for network with more anchor nodes and more uniform distribution. When density of anchor node reduced, localization performance of algorithm one based on range-free model is not very good. Then based on range-based model, algorithm two has been proposed. It is a localization algorithm for the network with fewer anchor nodes and also some unknown nodes which couldn’t locate themselves. Algorithm two named as Euclidean multilateral weighted centroid localization algorithm has integrated multilateral centroid localization algorithm with weighted centroid localization algorithm and Euclidean localization algorithm. It reduces node’s localization error and increases localization ratio based on range-based model. According to deployment of nodes in network, users can integrate algorithm one with algorithm two or choose only one localization algorithm between algorithm one and algorithm two.