The Research And Improvement Of DV-Hop Algorithm In Wireless Sensor Network

Network node localization is one of the most important technologies in Wireless SensorNetworks(WSN), when the event occurs, the sensor nodes must firstly know its ownlocation information in order to provide users the accurate information to take timely andeffective measures, thus obtaining accurate location information of nodes is especiallyimportant for the applications of wireless sensor network in the real world.In the current localization algorithm, Range-based localization algorithm andRange-free localization algorithm are widely used. although the range-based algorithm hassmall positioning errors, it requires hardware devices to assist positioning, resulting in highpositioning costs and big energy consumption. Range-free localization algorithm isprimarily based on the network connectivity, it does not need to measure the distance orrelative angle between the target node and the unknown node, the range-free algorithm hassimple positioning process and low positioning cost, what is more, it does not need thehardware to assist, but its positioning error is large.The paper describes several typical localization algorithms in wireless sensor network,and focuses on the DV-Hop algorithm. The paper proposes an improved DV-Hop algorithmto reduce the positioning error of nodes on the basis of analysis of error sources, the detailsare as follows:(1) The paper analyses the errors of DV-Hop algorithm, In case that the networknodes topology is uneven, the unknown nodes using the same average hop distance willlead to large errors between the hop distance and the actual distance,(2) The paper proposes an improved DV-Hop algorithm which weights the averagehop distance and the local average hop distance to correct the average hop distance ofunknown nodes, and the method improves the positioning accuracy of the algorithm.(3) In the third stage of DV-Hop algorithm, in order to reduce the errors whencomputing the coordinates of unknown nodes, we present an improved maximumlikelihood estimation method, furthermore, we revise the estimated coordinates ofunknown nodes. Comparing the improved maximum likelihood estimation of the paperwith the traditional maximum likelihood estimation and the two-dimensional hyperbolicpositioning algorithm, through the theoretical analysis we can proof that the improved maximum likelihood algorithm produces the least error when calculating the coordinates ofnodes, and the estimated coordinates of nodes is close to the actual coordinates.(4) The simulation of the improved algorithms and traditional DV-Hop algorithm byMatlab simulation software is conducted in the respect of positioning error、positioningcoverage and positioning effect. The simulation results show that the improved algorithmhas good positioning performance. Finally, the paper analyzes the communicationoverhead of the DV-Hop algorithm and the improved algorithm.