Study On Localization Algorithm Based On Range-free Technology For Wireless Sensor Networks

Wireless sensor networks are in the rapid development in recent years. Due to its greatpracticability and application value, wireless sensor networks have received great attention innumbers of application fields. And positioning technology is one of the supportingtechnologies of wireless sensor networks. In most cases, data form nodes that can’t confirmits location information has no practical significance. And the nodes’ hardware configurationwill be affected by many factors such as environment, power consumption, network area, andso on. It is unrealistic to configure location device or distance measuring equipment. So theresearch on localization algorithms based on the range-free technology has vital significance.This paper introduces the basic structure and characteristics of wireless sensor networks.And mainly analyze the localization technology in wireless sensor networks. At the same time,this paper introduces the classification of the localization technology, the basic algorithms andbasic calculation methods of positioning. Also, this paper states the various evaluationindexes for positioning effect. And do the simulation of centroid algorithm and APITalgorithm that are the classic ones bases on the range-free technology to compare. It analyzesthe positioning error and coverage rate in cases with obstacles or without, and details thecharacteristics and application fields of the localization algorithms combined with severalperformance indexes.Firstly, this paper analyzes and improves the APIT algorithm. APIT algorithm has highlocation accuracy, and low hardware requirements. But it needs to have a number of anchornodes as positioning reference. For unknown nodes that have neighbor anchor nodes fewerthan three or have no triangle that contains it, the localization can’t be done by APIT.According to the problem above, this paper uses the cycle positioning. It divided thelocalization into several times and just chooses the nodes that meet the requirements eachtime. During the positioning, the backtracking algorithm will be added in order to decreasemiscalculation and ensure the position precision. Then put the unknown nodes which havebeen located as anchor nodes to participate in the next round of localization to increase thecoverage rate.Then, this paper improves the DV-Hop algorithm. The algorithm is simple and put thehops between the nodes and the actual distance between anchor nodes as the positioning basis.But as using the hop distance as the linear distance between nodes, a certain number ofestimation error will come into being when estimate the distance between the unknown nodeand anchor node. This paper introduces the ideal distance between anchor nodes in order toexclude the data that has too much deviation with it when the average one hop distance iscalculated. This will decrease the impact of the error of average one hop distance in the later positioning. After the positioning, the coordinate correction will be added. It will use the errorfrom the positioning that use the located unknown node to corrected the actual coordinate ofthe nodes.Finally, this paper put forward an improved DV-Hop algorithm based on the RSSIcorrection, because that error will be produced when locate just rely on fixing connectedrelationship. It redefined the Hop count between nodes and joined the constraint conditions inthe Coordinate calculation process. It changes the calculation problem into an Optimizationproblem. Because the RSSI is unnecessary to change into precise distance, there will alsoobtain good results when the RSSI measurement has a certain error.