Underground Wireless Sensor Network Localization Algorithm Research

Wireless sensor network is a new multi-disciplinary overlapping researchfield emerging with constantly developing sensor technology, wirelesscommunication technology, information processing technology and varyingsocial demands. Node localization technology which is one of key supportingtechnologies of wireless sensor network has become a hot issue in wirelesssensor network.First of all, the principle and characteristics of several typical localizationalgorithm of wireless sensor network were introduced. Performance evaluationsof localization algorithm were also described. The characteristics of RSSIlocalization algorithm and its deficiencies when it was applied to personnellocalization of coal mime underground were argued, on the basis of study of theresearch status of personnel localization of coal mine underground, thecharacteristics of personnel positioning system based on wireless sensor networkand the special environment of coal mine. An improvement RSSI localizationalgorithm which takes advantage of the information of beacon node and takes environment factors into account was proposed. In order to exclude inaccurateRSSI values, the improved algorithm broadcast the information of ID,coordinate and hop of beacon node to the network just like DV_HOP algorithmdoes. The information of coordinate and hop of beacon node were used to fixposition error, so as to improve the anti-interference and positioning accuracy ofRSSI localization algorithm.Then, Aiming at the problem of the positioning result of sensor node is notas accurate, by the complex environment factors of underground coal mineroadway，a GIS-ANN correction method was proposed. The method utilizedmap management functions of GIS system and the characteristic that is insteadof having to establish the mathematical model of Artificial Neural Network. Thecorrection method consists of three parts: first, the establishment and training ofartificial neural network, the training sample sets made up of the experimentaldata and actual data; Second, to find the data which need to modify, throughmap match between positioning data and GIS map; Third, to fix the error ofpositioning data by trained artificial neural network. Experimental results showthat the method effectively fixed the error of the node coordinates, so as to improve the accuracy of the node positioning itself.Finally, according to the localization algorithm and localization correctionmethod proposed in this paper, an underground positioning system of wirelesssensor network was designed and implemented. Hardware parts were composedof Mote sensor nodes of Crossbow Company and soft parts were developedunder ArcGIS Server9.3platform of ERSI Company in the positioning system.Localization experiment which simulated mine coal underground wascommitted. The experimental results show that the WSN positioning system canmeet the demand of underground coal mine positioning.