Research On Localization Algorithm For Wireless Sensor Networks

Wireless sensor networks(WSN)is a kind of distributed network,which consists of a large number of small,low-cost sensor nodes with multi-function and multi information acquisition ability.They can communicate over a wireless link.The application prospect of wireless sensor network is very wide,it plays an irreplaceable role in many fields and has great development potential.The wide application of this technology will have a great influence on modern military,modern information technology,modern manufacturing industry and many important social fields.For most practical applications,if the exact location of nodes is not obtained,the collected data information is meaningless.The accurate location of nodes is a very important part of wireless sensor networks,so this dissertation proposes two improved algorithms to improve the location accuracy of traditional DV-Hop algorithm.The HTDV-Hop algorithm is based on the similarity factor fitting to improve the DV-Hop localization algorithm.The EG-DV-Hop algorithm based on machine learning and grey wolf optimization is proposed to solve the localization error caused by the limited number of anchor nodes and the average jump distance.HTDV-Hop,a location algorithm based on similarity factor and fitting function,is proposed in this dissertation.The algorithm is divided into four steps: the first step is to obtain the hop information from anchor nodes to all nodes;the second step is to calculate the similarity factor,and make a linear fitting between the similarity factor and the actual distance of the anchor node,using the fitting function to modify the distance between the anchor node and all the hop nodes;the third step is to correct the distance between anchor node and all the multi hop nodes;the last step is to use the improved distance to calculate the coordinates of unknown nodes using the least square method.The simulation results show that the location accuracy is improved by about 35.6% compared with the DV-Hop algorithm.On the other hand,an algorithm based on machine learning and grey wolf optimization(EG-DV-Hop)is proposed to increase the number of anchor nodes by virtual anchor nodes.EG-DV-Hop algorithm is divided into four steps: the first step is to forecast the unknown nodes' physical location by using the ELM learning model;the second step is to upgrade unknown nodes to virtual anchor nodes with the help of derivative nodes;third step is to modify the average hop distance with the grey wolf optimizer algorithm;the last step is to use the improved average hop distance to locate other unknown nodes.The simulation results show that compared with the DV-Hop algorithm,EG-DV-Hop method increases the localization accuracy by 25%.Through a large number of simulation experiments,the results show that,compared with other algorithms,the location algorithm based on similarity factor and fitting function(HTDV-Hop)and the location algorithm based on machine learning and grey wolf optimization(EG-DV-Hop)proposed in this dissertation not only have simple algorithms but also can achieve high-precision location,which has a good application prospect.