基于经验模态分解的室内指纹定位算法

Indoor fingerprint localization algorithm based on empirical mode decomposition

WiFi信号的波动问题是影响指纹定位精度的主要因素之一,针对该问题设计了基于经验模态分解(EMD)的室内指纹定位算法,利用EMD方法在保持信号原有特征的同时实现了平滑噪声的目的。同时,针对传统的加权K最邻近(WKNN)算法在参考点数量大、指纹特征维度高时存在的计算量大、定位实时性低的问题,采用Ball Tree的近邻检索方法优化指纹匹配阶段的数据查询结构,提高了指纹匹配的速度。实验结果表明:相较于常用的传统WiFi指纹定位算法,所提基于经验模态分解的室内指纹定位算法可获得17%以上定位误差的改善,同时指纹匹配耗时减少了45%左右,有效提高了定位的精度与实时性。

The fluctuation problem of WiFi signal is one of the main factors affecting the accuracy of fingerprint localization.In order to solve the problem,an indoor fingerprint localization algorithm based on Empirical Modal Decomposition(EMD)was designed,and the EMD method was used to achieve the purpose of smoothing the noise while maintaining the original characteristics of the signal.At the same time,aiming at the problems of large amount of calculation and low real-time localization of the traditional Weighted K-Nearest Neighbor(WKNN)algorithm when the number of reference points is large and the fingerprint feature dimension is high,the Ball Tree’s nearest neighbor retrieval method was used to optimize the data query structure in the fingerprint matching stage and improve the speed of fingerprint matching.The experimental results show that,compared with the commonly used traditional WiFi fingerprint localization algorithm,the proposed indoor fingerprint localization algorithm based on empirical mode decomposition can obtain more than 17%improvement in localization error while the fingerprint matching time is reduced by about 45%,which effectively improves the accuracy and effectiveness of localization.