基于克里金空间插值的位置指纹数据库建立算法*

为解决室内定位系统中建立位置指纹数据库工作量庞大的问题，提出了一种融合信号衰减因素的普通克里金空间插值建库算法。该方法首先采用高斯滤波对有限预选参考点的信号强度采样数据进行预处理，并基于球状模型由参考点预处理数据拟合出空间变异函数，然后采用普通克里金插值法对其它位置的信号强度进行估值并生成相应的位置指纹，最后由有限实测数据生成大容量高分辨位置指纹数据库，并通过KNN_Filter算法和对数权重算法完成了定位仿真，验证了该建库方法的有效性。仿真实验结果表明，该算法在保证定位精度的前提下相比传统建库方法可降低40%左右的工作量，提高了室内位置指纹定位方法的工作效率。

Location fingerprint database construction algorithm based on Kriging spatial interpolation

In order to solve a problem about the huge workload of established fingerprint database in the indoor positioning system, this thesis proposed an ordinary Kriging spatial interpolation algorithm fused signal attenuation factor. Firstly, this method used Gaussian filter to deal with the sampling data of the limited preselected reference node signal strength, then obtained the spatial variation function by fitting the obtained reference data with the spherical model. Then it used ordinary Kriging spatial interpolation to acquire the estimation of other node signal strength and generated location fingerprint. Finally, it could generate a large capacity and high resolution position fingerprint database by finite field measurement data. This paper completed positioning simulation by the KNN_Filter algorithm and logarithm weighting algorithm to prove the validity of this established database method. The simulation results show that compared with the traditional established database method, the algorithm reduces the workload of about 40% under the premise of ensuring accuracy of the positioning and improves working efficiency of the indoor location fingerprint positioning method.