通信学报 ›› 2015, Vol. 36 ›› Issue (6): 13-21.doi: 10.11959/j.issn.1000-436x.2015147

• 学术论文 • 上一篇    下一篇

带有同步预测的WBAN时序数据融合算法

王汝言,翟美玲,吴大鹏   

  1. 重庆邮电大学 宽带泛在接入技术研究所,重庆 400065
  • 出版日期:2015-06-25 发布日期:2017-05-11
  • 基金资助:
    国家自然科学基金资助项目;重庆市自然科学重点基金资助项目;重庆市自然科学重点基金资助项目;重庆邮电大学青年自然科学基金资助项目;重目庆市青年科技人才培养计划基金资助项目

Time series data aggregation algorithm with synchronous prediction for WBAN

Ru-yan WANG,Mei-ling ZHAI,Da-peng WU   

  1. Broadband Ubiquitous Network Research Laboratory,Chongqing University of Posts and Telecommunications,Chongqing 400065,China
  • Online:2015-06-25 Published:2017-05-11
  • Supported by:
    The National Natural Science Foundation of China;The Natural Science Foundation of Chongqing;The Natural Science Foundation of Chongqing;The Foundation of Chongqing University of Posts and Telecommunications;Youth Talents Training Project of Chongqing Science & Technology Commission

摘要:

提出一种带有同步预测的时序数据融合算法,利用多分辨率分析特性对采集的原始数据进行预处理,挖掘反映人体生理状态的本质特征,进而采用同步预测机制在感知节点和汇聚节点处分别建立轻量级预测模型,消除网内冗余数据的传输以降低能耗。结果表明所提出的融合算法具有较高的预测精度,能够实现低开销的无线体域网时序数据融合。

关键词: 无线体域网, 数据融合, 多分辨率分析, 最小二乘支持向量机, 同步预测机制

Abstract:

Due to the nonlinearity and nonstationarity of the physiological data sensed by WBAN,data aggregation cannot be effectively achieved according to the time-domain trend of data.Therefore,a novel time series data aggregation algorithm with synchronous prediction was proposed.By preprocessing the original sensing data with the multi-resolution analysis,the inherent characteristics of the physiological data can be obtained to establish a light-weight synchronous prediction model at both the sensor and sink.Numerical results show that the proposed aggregation algorithm can achieve a favorable prediction precision and a low energy consumption rate by eliminating the in-network data redundancy.

Key words: wireless body area network, data aggregation, multi-resolution analysis, LSSVM, synchronous prediction mechanism

No Suggested Reading articles found!