基于SVR-Kriging插值的矿井工人二维指纹定位数据库构建算法

发布时间：2019-04-10 07:35

【摘要】：为突破矿井工人指纹定位中1维模型在定位精度上的局限性,该文提出一种矿井工人2维指纹定位数据库构建算法,并通过SVR-Kriging插值法解决因2维模型带来的数据采集工作量大的问题。首先,通过高斯滤波对采集的采样点位置指纹信息进行预处理,并利用支持向量回归由采样点数据拟合变异函数。然后采用Kriging插值法补全2维网格划分中的未采样区域的位置指纹信息。最后综合采样点与插值点的位置指纹信息建立矿井工人指纹信息数据库,为后续矿井工人指纹定位奠定基础。仿真结果表明,该文算法在减少数据采集工作量的同时保证了算法的可行性与有效性,且在进行位置指纹定位时能够保证较高的精度。
[Abstract]:In order to break through the limitation of 1-D model in location accuracy of mine workers' fingerprint, this paper presents an algorithm for constructing 2-D fingerprint location database of mine workers. SVR-Kriging interpolation method is used to solve the problem of large data collection workload caused by 2-D model. Firstly, the position fingerprint information of sampling points is pre-processed by Gaussian filtering, and the variance function is fitted by support vector regression (SVR) from sampling point data. Then the Kriging interpolation method is used to complete the position fingerprint information of the unsampled area in the 2-D mesh generation. Finally, the database of mine workers' fingerprint information is established by integrating the location fingerprint information of sampling point and interpolation point, which lays a foundation for the subsequent location of mine worker's fingerprint. The simulation results show that the proposed algorithm not only reduces the workload of data collection but also ensures the feasibility and effectiveness of the algorithm, and the accuracy of location fingerprint location can be guaranteed.
【作者单位】：电子工程学院;
【基金】：国家自然科学基金(61273302)~~
【分类号】：TP311.13

【相似文献】