基于融合聚类的地磁室内定位算法研究
首发时间:2022-03-16
摘要:针对地磁室内定位的实用性、方便性、以及准确性,本文提出了密度峰值算法(DPCA)和K均值算法(K-means)融合的聚类算法以及使用标准化欧式距离对动态时间规整算法(DTW)进行改进。K-means算法难以确定K值,DPCA算法可以利用决策图得出聚类中心,继而得到K值,提高了聚类效果。聚类结束后在线阶段使用DTW算法匹配,在DTW算法中计算两个地磁指纹之间的距离传统的是使用欧式距离,本文采用标准化欧式距离克服了欧氏距离各个分量都一样的缺点。聚类算法和匹配算法的改进,提高了地磁室内定位的精度。
关键词: 室内定位 K均值聚类 密度峰值聚类算法 动态时间规整算法
For information in English, please click here
Research on geomagnetic indoor location algorithm based on fusion clustering
Abstract:Aiming at the practicability, convenience and accuracy of geomagnetic indoor positioning, this paper proposes a clustering algorithm based on the fusion of density peak algorithm (DPCA) and K-means algorithm (K-means), and improves the dynamic time warping algorithm (DTW) with standardized Euclidean distance. K-means algorithm is difficult to determine the K value. DPCA algorithm can use the decision map to get the clustering center, and then get the K value, which improves the clustering effect. After clustering, the online phase uses DTW algorithm to match. In DTW algorithm, the distance between two geomagnetic fingerprints is calculated. Traditionally, the Euclidean distance is used. In this paper, the standardized Euclidean distance is used to overcome the disadvantage that each component of Euclidean distance is the same. The improvement of clustering algorithm and matching algorithm improves the accuracy of geomagnetic indoor positioning.
Keywords: Indoor positioning K-means clustering Density peak clustering algorithm Dynamic time warping algorithm
基金:
引用
No.****
动态公开评议
共计0人参与
勘误表
基于融合聚类的地磁室内定位算法研究
评论
全部评论0/1000