室内电磁精确定位系统校准技术的研究

发布时间：2018-06-01 17:28

本文选题：电磁定位 + 多项式拟合　；参考：《哈尔滨工程大学》2016年硕士论文

【摘要】：虚拟现实是一项人机交互式界面技术,可以较为真实的模拟人处于自然环境中的行为。空间位置跟踪技术是虚拟现实技术中重要的组成部分,其性能表现直接影响虚拟现实交互过程中的真实感。由于电磁定位系统具有精度良好、实时性高、工作范围适中等优点,使之成为了室内小范围虚拟现实系统实现的技术首选。电磁定位系统的磁场分布易受外界环境干扰而发生形变,致使定位结果存在较大的误差,使用前需要对电磁定位系统进行校准。本文采用多项式拟合和邻点插值方法,对室内电磁定位系统的误差进行校正,校正后的结果都达到了实际应用需求。多项式拟合方法的前期样本数据采集过程较快,校正过程计算量相对较小,但精度相对不高;邻点插值方法的前期样本数据采集过程较慢,校正过程计算量较大,但取得了更高的精度。本论文具体工作归纳如下:首先,通过对现有室内定位系统及其优缺点分析,论证了选用电磁定位技术是解决室内三维位置跟踪技术的有效手段。构建了实际测试实验环境,设计了定位系统性能测试实验方案,分析了电磁定位系统磁场空间存在的非线性形变及其引起的误差。其次,提出了一种基于多项式拟合的室内电磁定位系统全局误差校正方法。设计了建立多项式拟合关系所需样本数据及测试数据采集方案,为了消除样本数据存在的扭转形变,提出了采用旋转变换矩阵对样本数据进行预处理。通过最小二乘合法拟合样本数据的圆曲线,并提取拟合后的圆心及半径坐标,建立了电磁坐标数据库。采用多项式拟合法分别在水平方向和垂直方向上建立电磁定位数据与真实空间定位数据之间的映射关系。对测试数据进行处理,验证了该校正算法的精度和有效性。最后,提出了一种基于邻点插值的误差校正方法。设计了邻点插值所需样本数据采集方案,提出了采用Delaunay三角剖分法对电磁定位采样数据进行四面体结构划分,将待校正电磁空间划分成若干个四面体集合,随后对位于不同四面体内的待校正点进行局部校正。为了快速定位待插值点所在四面体,提出了利用八叉树的方法对四面体集合建立八叉树索引结构。提出采用重心坐标加权插值法对待插值点进行误差计算,提高校正精度。对测试数据进行处理,验证了该校正算法得到了更高的精度。
[Abstract]:Virtual reality is a human-computer interactive interface technology, which can simulate the behavior of human being in the natural environment. Spatial position tracking technology is an important part of virtual reality technology, and its performance directly affects the reality of virtual reality interaction process. Because the electromagnetic positioning system has the advantages of good precision, high real-time performance and moderate working range, it has become the first choice of technology for the realization of indoor small-scale virtual reality system. The magnetic field distribution of the electromagnetic positioning system is easily distorted by the external environment, which results in a large error in the positioning results. It is necessary to calibrate the electromagnetic positioning system before use. In this paper, polynomial fitting and neighbor interpolation are used to correct the errors of indoor electromagnetic positioning system. The previous sample data acquisition process of polynomial fitting method is relatively fast, the calculation of correction process is relatively small, but the accuracy is relatively low, the early sample data acquisition process of neighbor interpolation method is slow, and the calculation of correction process is large. But higher accuracy has been achieved. The specific work of this paper is summarized as follows: firstly, through the analysis of the existing indoor positioning system and its advantages and disadvantages, it is demonstrated that the electromagnetic positioning technology is an effective means to solve the indoor three-dimensional position tracking technology. In this paper, an actual testing environment is constructed, and the performance test scheme of the positioning system is designed. The nonlinear deformation in the magnetic field space of the electromagnetic positioning system and the error caused by it are analyzed. Secondly, a global error correction method for indoor electromagnetic positioning system based on polynomial fitting is proposed. In order to eliminate the torsional deformation of the sample data, the rotation transformation matrix is proposed to preprocess the sample data. The circle curve of the sample data was fitted by least square method, and the fitted center and radius coordinates were extracted, and the electromagnetic coordinate database was established. The mapping relationship between electromagnetic positioning data and real spatial positioning data is established in horizontal and vertical directions by polynomial fitting method. The accuracy and validity of the proposed algorithm are verified by processing the test data. Finally, an error correction method based on neighbor interpolation is proposed. A sample data acquisition scheme for neighbor interpolation is designed, and a tetrahedron structure partition of electromagnetic location sampling data by Delaunay triangulation method is proposed. The electromagnetic space to be corrected is divided into several tetrahedron sets. Then local correction is made for the points to be corrected at different quadrihedral bodies. In order to locate the tetrahedron of the point to be interpolated quickly, an octahedral index structure for tetrahedron set is proposed by using octree. In order to improve the accuracy of correction, a weighted interpolation method for barycentric coordinates is proposed to calculate the errors of interpolation points. The processing of test data verifies that the proposed algorithm has higher accuracy.
【学位授予单位】：哈尔滨工程大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.9

【参考文献】