融合反距离加权和矩阵填充的位置指纹库构建

为解决位置指纹定位在离线阶段构建位置指纹库时耗费的人力和时间成本较大，构建指纹库效率低和利用空间插值法构建的指纹库精度不高的问题，提出了一种融合反距离加权和矩阵填充的位置指纹库构建算法。该算法仅需人工采集定位区域内少量参考点的接收信号强度值用作信标点指纹信息，结合反距离加权算法特性计算出次信标点指纹信息，根据位置指纹库数据矩阵的低秩性，应用奇异值阈值矩阵填充算法构建出位置指纹数据库。仿真实验结果表明，所提算法有效降低了矩阵填充算法构建位置指纹库所需的人工和时间成本，构建出的位置指纹库定位性能优于反距离加权和克里金空间插值法，接近传统人工采集法，显著地提高了位置指纹库的构建效率。

Construction of location fingerprint database integrating inverse distance weighted and matrix completion

To solve the problems of constructing location fingerprint database on offline stage of location fingerprint positioning,including large manpower and time cost,the low efficiency of constructing location fingerprint database and the low precision of fingerprint database constructed by spatial interpolation method,a location fingerprint database construction algorithm integrating inverse distance weighted(IDW)and matirx completion is proposed.Only the received signal strength values of a few reference points in the location area are needed to collect manually to use as the fingerprint information of beacon points .Then the fingerprint information of secondary beacon points is calculated by using IDW algorithm.Finally,the location fingerprint database is constructed by using singular value thresholding(SVT)matrix completion algorithm according to the low rank property of data matrix in location fingerprint database.The simulation results show that the proposed algorithm greatly reduces the manual and time cost of constructing the location fingerprint database by matrix completion algorithm.The positioning performance of the constructed location fingerprint database is better than that of IDW and Kriging spatial interpolation method,which is close to the traditional manual acquisition method.The proposed algorithm greatly improves the efficiency of constructing location fingerprint database.