Research On The Calibration Technology For Indoor EM Precise Positioning System

Virtual reality is an interaction technology for human-computer interface, which can imitate a person's behavior in the natural environment by simulation system. Space position tracking technology is important for virtual reality because it has directly affection on the performance of the virtual reality. Since electromagnetic positioning system has high accuracy,high real-time and moderate operation range, it is the first choice for indoor virtual reality in small region. However, the field distribution of positioning system is vulnerable to external interference caused by deformation which leads to large errors.In this paper,we use the methods of polynomial fitting and adjacent point interpolation to calibrate the indoor electromagnetic positioning system. Polynomial fitting has the merits of faster pre-sample data acquisition, less calculation and has the shortcoming of low precision. Adjacent point interpolation has the shortcomings of slower pre-sample data acquisition, more calculation, but has the merit of high precision.First, it's investigated that existing indoor positioning technologies. The nonlinear deformation and error of the electromagnetic positioning system are analyzed. Both the environment and system performance experiments are carried out which are satisfied with the practical application requirements.Second,it's proposed an error correction method of indoor electromagnetic positioning system global error based on polynomial fitting. The sample data and the test data acquisition scheme are designed to establish the polynomial fitting relation. To eliminate the torsion deformation of the sample data, the sample data was pre-treated using the rotation transformation matrix. To establish the electromagnetic coordinate database, the center and radius of each circle of the coordinate fitting are extracted by the least squares fit circle curve of sample data,. The polynomial fitting method is adopted to establish the relationship between the electromagnetic positioning data and the real spatial location data in the horizontal and vertical directions. The test result has verified the accuracy and effectiveness of the proposed algorithm.Finally, it is proposed an error correction method based on neighbor interpolation. The sample data acquisition scheme is designed. The Delaunay triangulation method is used to divide the electromagnetic positioning data into the tetrahedral structure. The correction of electromagnetic space is divided into several tetrahedron sets, then the calibration is realized in different calibration tetrahedron. To determine which tetrahedral the point locates, octree is adopted to establish octree index structure of the tetrahedron set. To improve the correction precision, the barycentric coordinates weighted interpolation method is adopted to calculate the error of the interpolation point. The test result showed that the correction algorithm is more accurate.