Font Size: a A A

Indoor Localization Based On RF Fingerprinting Method

Posted on:2015-03-16Degree:MasterType:Thesis
Country:ChinaCandidate:P J DuanFull Text:PDF
GTID:2268330428977794Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
As one of the supporting technologies of wireless sensor network,nodepositioning technology can help user to control network topology, routing,monitoring and so on. In the complex indoor environment, the radio signalpropagation will produce multipatheffect. This factor will have a great effect onthe positioning results. Compared with the outdoor positioning technology,indoor localization technology has its particularity and complexity. Therefore,study on indoor localization technology in wireless sensor network hasimportant theoretical significance and practical value.Radio-map locating mathod uses signal strength informations of multiplereceiver locations to positioning the user’s coordinates, and the process isempirically-determined and theoretically-computed. This method consists ofoff-line training stage and on-line locating stage. And the characteristic is doesnot require additional hardware support, wide coverage, no line-of-sight distancerequirements, low cost, etc. This method has become the mainstream of theindoor positioning technology. In this paper, we present the research of indoorlocalization method based on radio-map:During offline-training stage, we suggest using movement track chain tobuild Radio-map in order to solve the huge training intensity, high couplingdegree between the radio-map and environment, and the reconstruction problem.The method records the received signal strength and location coordinates in themoving process in order to construct the radio-map, to avoid the largepositioning error problem between the method of interpolation algorithm andthe method of model formula. As result it can construct the radio-mapconveniently and fastly in the indoor environment. In the online-positioning stage, we present an indoor mobile nodelocalization algorithm based on finite state machine to resolve the low precisionproblem of the traditional positioning method, and the complex problem ofprobabilistic localization algorithm. The algorithm depend on state transition offinite state machine, use current received signal strength indicator and combinedwith historical data to constraints the moveable space of the unknown target inorder to eliminate position that can’t be appearance during moving process, andimprove the positioning accuracy.The experiments results show that, using movement track chain to buildRadio-map can greatly reduce the work intensity of off-line training almostwithout affect the positioning accuracy. And the localization algorithm based onfinite state machine estimated location more precise. In the areas of15.526.99m2, the average positioning accuracy of the algorithm is1.194meters.
Keywords/Search Tags:wireless sensor network, indoor wireless position estimate, radio-map, finite state machine, Received Signal Strength Indicator (RSSI)
PDF Full Text Request
Related items