无线传感网中一种基于自回归模型的数据收集方案  

A DATA COLLECTION SCHEME IN WIRELESS SENSOR NETWORKS BASED ON AUTO REGRESSIVE MODEL

作　　者：李余琪[1] 甘金明[1] 周顺先[1] 

机构地区：[1]梧州学院计算机系,广西梧州543002

出　　处：《计算机应用与软件》2014年第3期99-103,共5页Computer Applications and Software

基　　金：国家自然科学基金项目(51074097);广西重大科技攻关项目(桂科攻12118017-6);广西自然科学基金项目(2011GXNSFA018165);广西教育厅科研重点项目(201102ZD034)

摘　　要：数据收集问题是无线传感网中的一个研究热点。针对现有数据收集方法的不足,提出一种基于自回归模型的数据收集方案。首先分析感知数据稀疏性变化情况对于重构性能的影响,然后基于自回归模型对压缩感知重构问题进行建模,最后sink利用时间相关性来对重构误差进行评价,并根据重构误差要求来决定是否需要增加测量次数,从而实现对感知数据的自适应重构。仿真实验结果表明,该方法是有效的,在数据重构精度以及网络生命周期等方面要优于传统的方法。Data collection is a hot topic in wireless sensor networks currently. Aiming at the disadvantage of existing data collection methods, we propose an AR model-based data collection scheme. First, the scheme analyses the impact of sparsity variation of sensitive data on reconstruction performance, and then models the compressed sensing reconstruction based on the AR model, finally, the sink evaluates the reconstruction error using the temporal correlation, and decides whether or not to increase the times of measurements according to the requirements of reconstruction error, so as to realise the adaptive reconstruction of sensing data. Simulation experimental results show that our method is effective, and is superior to traditional methods in terms of the data reconstruction accuracy and the lifecycle of network.

关 键 词：无线传感网 数据收集 自回归模型 测量次数 重构精度 网络生命周期 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]