基于FCM聚类及位置区切换室内地磁定位研究.docVIP

基于FCM聚类及位置区切换室内地磁定位研究 　　摘 要： 针对目前室内地磁定位技术中存在的地磁指纹库太大及地磁指纹不惟一所造成的运算量大、定位误差大的问题，提出一种基于模糊C?均值聚类与位置区切换相结合的室?鹊卮哦ㄎ环椒āＪ紫榷源?定位点的位置区进行判别，然后对该点地磁矢量所属的聚类进行判别，最后在其所属的位置区及聚类对应的匹配范围内进行匹配定位。经过以上两步判别，可以缩小地磁指纹的匹配范围，减小匹配运算量，同时降低了地磁矢量接近但地理位置上相距较远时加大定位误差的风险。经实验及仿真验证，改进的方法相比WKNN（KNN）算法减少了约41%的运算耗时（1 000个地磁指纹的情况下），定位精度提高了约76%。 　　关键词： 室内地磁定位； 地磁定位技术； 地磁指纹库； 模糊C?均值聚类； 位置区切换； 匹配定位 　　中图分类号： TN953+.7?34； TP391 文献标识码： A 文章编号： 1004?373X（2018）14?0096?05 　　Research on indoor geomagnetic positioning based on FCM 　　clustering and location area switching 　　SONG Yu， YU Wenju， CHENG Chao 　　（School of Computer Science and Engineering， Changchun University of Technology， Jilin 130012， China） 　　Abstract： In allusion to the problems of big amount of computations and big positioning errors caused by much hugeness of the geomagnetic fingerprint database and non?uniqueness of the geomagnetic fingerprints existing in the current indoor geomagnetic positioning technology， an indoor geomagnetic positioning method based on the combination of fuzzy C?means clustering and location area switching is proposed. The location area of the to?be positioned point is discriminated， the cluster that the geomagnetic vector of the point belongs to is discriminated， and the matching positioning is performed in the location area of the point and within the matching range corresponding to the cluster. The above two discriminations can reduce the matching range of geomagnetic fingerprints， the amount of matching computations， and the risk of increasing positioning errors when the geomagnetic vectors are approximate but geographically far away. The results of the experiment and simulation verification show that in comparison with the WKNN（KNN） algorithm， the improved method can reduce the computation time consumption by about 41% in the case of 1000 geomagnetic fingerprints and improve the positioning accuracy by about 76%. 　　Keywords： indoor geomagnetic positioning； geomagnetic positioning technology； geomagnetic finger