摘要
无线传感器网络中节点监控时采集的数据具有时间和空间上的相关性,给节点通信带来负担,缩短网络生命周期。为降低冗余数据,提出一种基于NARX神经网络的分簇数据融合算法(N-CDAA)。将NARX神经网络时序预测模型和基于矢量量化的分簇路由协议有机结合,从时间和空间相关性上消除冗余,把融合后的少量数据发送给汇聚节点,提高数据收集效率,延长网络生存时间。实验结果表明,该算法预测精度高,可有效降低数据传送量,到达延长网络生命周期的目的。
In wireless sensor networks(WSNs),the data collected by the nodes have temporal-spatial correlation,which brings the burden to the nodes’communication and shortens the network life cycle.To reduce redundant data,a clustering data fusion algorithm based on NARX neural network(N-CDAA)was proposed.The NARX neural network time series prediction model was combined with the clustering routing protocol based on vector quantization,eliminating the redundancy of temporal-spatial correlative.A small amount of fused data was sent to the Sink node.The proposed algorithm can improve data collection efficiency and prolong the network survival time.Simulation results show that the algorithm has high prediction accuracy,and it can effectively reduce the data transfer,and achieve the purpose of prolonging the network life cycle.
作者
范时平
何超杰
FAN Shi-ping;HE Chao-jie(College of Communication and Information Engineering,Chongqing University of Posts and Telecommunication,Chongqing 400065,China)
出处
《计算机工程与设计》
北大核心
2018年第4期938-942,共5页
Computer Engineering and Design
基金
重庆高校优秀成果转化基金项目(KJZH17116)
重庆市教委科学技术研究基金项目(KJ1400422
KJ1500441
KJ1400431)
重庆市科委重点产业共性关键技术创新专项基金项目(CSTC2015ZDCY-ZTZX40001)
关键词
无线传感器网络
数据融合
神经网络
预测模型
分簇
wireless sensor networks(WSNs)
data aggregation
neural network
prediction model
clustering
