The Research Of Indoor Geomagnetic Positioning Technology Based On Crowdsourcing

With the rapid development of mobile Internet,various mobile internet services have been deeply into our daily lives.With the development of the navigation,maps and other applications,LBS(Location Based Service)has been an indispensable basis for mobile internet applications in our lives.The GPS is main signal of outdoor positioning scenes at present.However,because the satellite signals is disturbed in indoor scenes,the traditional GPS positioning technology can't provide good service in indoor situation.These is a need to put forward an indoor positioning system of high accuracy and good reliability.Since the geomagnetic positioning has better signal stability and there are no need to deploy a dedicated infrastructure,the geomagnetic has been a top hot technology of indoor positioning.We are given many research results based on geomagnetic signals.Existing magnetic positioning system requires a lot of training of geomagnetic fingerprint by professionals.These systems have the shortage of high cost and low promotional,which restrict the further development of the geomagnetic positioning technology.Based on these shortage and study of crowdsoucing,this paper combined traditional geomagnetic positioning technology with the method of crowdsourcing.We put up with a low cost and adaptive training method of geomagnetic fingerprint.This paper designed a geomagnetic fingerprint collector based on turning recognition,initial positioning and walking direction identification on the android client.We split the geomagnetic fingerprint data into atom paths based on INS information,walking direction,step length and turn signs.In the server side,this paper attached coordinates to the geomagnetic fingerprint with the INS information.In order to correct the accumulated error in training stage owing to the INS,the server side make cluster and fusion operation with the atom paths.Finally,the server established a geomagnetic fingerprint database for localization algorithm.Meanwhile,this paper did the indoor topology mining.Based on the proposed indoor location algorithm of crowdsourcing,this paper designed the client Direction and the server CrowdSoucing.After large number of experiments and simulation,the geomagnetic fingerprint database turned out to be a low cost and high accuracy database.The proposed algorithm is a reliable.