基于分层的体域网异常数据检测方法
Hierarchical anomaly data detection method for body area networks
廖栋森 1祝长鸿 1余琪琦 1任君玉 2黄福莹 1覃团发1
作者信息
- 1. 广西大学计算机与电子信息学院,广西南宁 530004;广西大学 多媒体通信与网络技术重点实验室,广西 南宁 530004
- 2. 广西大学 多媒体通信与网络技术重点实验室,广西 南宁 530004
- 折叠
摘要
针对无线体域网(wireless body area network,WBAN)异常数据检测方法忽视人体异常数据的连续性,缺乏异常数据集检测等问题,提出一种基于Hampel滤波器和DBSCAN分层的 WBAN异常数据检测方法.根据时间相关性利用Hampel滤波器检测异常数据点,保证数据的连续性,使用改进的基于滑动时间窗的DBSCAN算法,检测异常数据集.实验结果表明,所提方法和其它方法相比,实现了分层的异常数据检测,在保证检测精度的同时准确标注出了异常数据集,具有空间复杂度小的优势.
Abstract
Aiming at the problems that the wireless body area network(WBAN)abnormal data detection method ignores the con-tinuity of human abnormal data and lacks abnormal data set detection,a hierarchical WBAN abnormal data detection method based on Hampel filter and DBSCAN was proposed.According to the time correlation,Hampel filter was used to detect abnor-mal data points to ensure the continuity of data,and the improved DBSCAN algorithm based on sliding time windows was used to detect abnormal data sets.Simulation results show that compared with other methods,the proposed method achieves hierarchical abnormal data detection,and accurately marks the abnormal data set while ensuring the detection accuracy.
关键词
无线体域网/异常检测/分层/聚类/连续性/时间窗/检测Key words
wireless body area networks/anomaly detection/hierarchy/DBSCAN/continuity/time window/detection引用本文复制引用
出版年
2024
NETL
NSTL
万方数据