无线传感网络多源数据特征融合方法研究  

Research on Multi-Source Data Feature Fusion Methodsfor Wireless Sensor Networks

作　　者：陈宏 蒋文贤[2] 黄丽萍 余翀翀 CHEN Hong;JIANG Wenxian;HUANG Liping;YU Chongchong(Department of IT Development and Management,Huaqiao University,Xiamen Fujian 361021,China;College of Computer Science and Technology,Huaqiao University,Xiamen Fujian 361021,China)

机构地区：[1]华侨大学信息化建设与管理处,福建厦门361021 [2]华侨大学计算机科学与技术学院,福建厦门361021

出　　处：《传感技术学报》2024年第12期2131-2136,共6页Chinese Journal of Sensors and Actuators

基　　金：福建省自然科学基金项目(2017J01776)。

摘　　要：在无线传感网络中,不同的传感器节点可能会收集到重叠或相似的数据,造成计算、存储和传输资源的浪费。为此,提出一种无线传感网络多源数据特征融合方法。结合互补集合经验模态分解和小波阈值去噪方法,在保留数据主要特征的同时去除噪声。通过主成分分析提取多源数据的第一层和第二层特征,并将其级联为最终提取的多源数据特征。采用模糊数学中的最大最小贴近度描述不同特征之间的距离,实现多源数据特征融合。仿真结果表明,所提方法应用后的节点死亡率低于5%,融合延迟小于4 ms,平均节点能耗保持在4.5 J以下,不同场景下的特征融合精度高于86.3%。表明所提方法具有良好的多源数据特征融合性能,可以有效地提高无线传感网络的数据传输性能。In wireless sensor networks,different sensor nodes may collect overlapping or similar data,resulting in waste of computing,storage,and transmission resources.To this end,a multi-source data feature fusion method for wireless sensor networks is proposed.By combining complementary set empirical mode decomposition and wavelet threshold denoising methods,noise is removed while retaining the main features of the data.The first and second layer features of multi-source data are extracted by using principal component analy-sis,which are cascaded into the final extracted multi-source data features.The maximum and minimum closeness degree in fuzzy mathe-matics is used to describe the distance between different features,achieving multi-source data feature fusion.The simulation results show that the node mortality rate after the application of the proposed method is less than 5%,the fusion delay is less than 4 ms,the average node energy consumption remains below 4.5 J,and the feature fusion accuracy in different scenarios is higher than 86.3%,indicating that the proposed method has good multi-source data feature fusion performance and can effectively improve the data transmission per-formance of wireless sensor networks.

关 键 词：无线传感网络 多源数据特征 特征融合 经验模态分解 小波阈值去噪 主成分分析 亲信度 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]