| With the increasing demand for navigation and location services in people's daily life,location-based service(LBS)is profoundly changing people's lifestyles.As a part of "last mile" of navigation and positioning,indoor positioning has important research value.The positioning technology based on MM(magnetic field matching)has great potential because environmental magnetic field is ubiquitous,which has attracted the attention of many scholars and enterprises.However,there are two difficulties in MM.Firstly,the construction efficiency of the magnetic fingerprint database and the magnetic dimension are difficult to guarantee at the same time,and the magnetic dimension is closely related to positioning accuracy;Secondly,the user or the integrated positioning system cannot be informed of the quality of MM positioning,that is,the integrity cannot be guaranteed.To solve these problems,this paper designs a high-dimensional magnetic fingerprint database construction scheme,a high-dimensional MM algorithm and studies the integrity monitoring method of the MM positioning system.The details are as follows.Firstly,a construction scheme of high-dimensional magnetic fingerprint database is designed based on Foot-INS scheme.Continuous,reliable foot position is obtained through Foot-INS scheme and projected to the handheld acquisition device(for example,a phone)because they have a fixed relative positional relationship at specific moments.The accuracy of position and attitude of the acquisition device is guaranteed by extended Kalman Filter algorithm,optimal smoothing algorithm and iterative method.The accuracy of magnetic data is ensured by magnetometer calibration and mean filtering method and the density of magnetic fingerprints is ensured by an interpolation method in the construction scheme.Experimental results show that the construction scheme can achieve 1?2dm position accuracy and 1?3deg attitude accuracy of the acquisition device,and the accuracy of repeated measurements of magnetic information in magnetic fingerprint database reaches 10-30 m Gauss.Secondly,taking advantage of high-dimensional magnetic information(projection components in north,east and vertical directions of magnetic vector),a highdimensional MM algorithm is designed based on a reliable high-dimensional magnetic fingerprint database.Dynamic time warping algorithm is used to estimate magnetometer bias because magnetometer bias doesn't affect the shape of the magnetic sequence in the body frame,which solves the problem of the dependence of magnetometer calibration in MM.Experimental results show that the online estimation accuracy of magnetometer bias is 10?30m Gauss.On this basis,taking advantage of high discrimination of high-dimensional magnetic information,there is a significant improvement on reliability and availability of MM and positioning accuracy of highdimensional MM algorithm is improved by about 25% compared with low-dimensional MM algorithm.Thirdly,an integrity monitoring algorithm based on magnetic features is designed and implemented.Through simulation and statistical analysis,the algorithm studies the relationship between magnetic features and integrity index "positioning theoretical standard deviation",and realizes real-time calculation of positioning theoretical standard deviation.Experimental results show that there is a certain linear relationship between magnetic features and positioning standard deviation for specific magnetic fingerprint database.However,the validity of positioning standard deviation isn't sufficient enough and the integrity monitoring algorithm needs to be further improved.In summary,this paper essentially upgrades the existing indoor MM positioning scheme with the construction scheme of high-dimensional magnetic fingerprint database and the high-dimensional MM algorithm,and fully explores the integrity monitoring algorithm,which makes technical preparations for the engineering and practicality of MM positioning. |
PDF Full Text Request