阴影模型的正则化无设备重建与实时定位  

Device-free Reconstruction and Real-time Location Based on Shadowing Model in Radio Tomographic Imaging

作　　者：熊一枫 卢继华[1] 何梓珮 曹晨曦[1] 

机构地区：[1]北京理工大学信息与电子学院,北京100081

出　　处：《自动化学报》2015年第6期1159-1165,共7页Acta Automatica Sinica

基　　金：国家高技术研究发展计划(863计划)(2012AA121604);国家自然科学基金(61002014;61101129;61227001;61072050)资助~~

摘　　要：在综合静态无线射频层析成像(Radio tomographic imaging,RTI)算法基础上,给出了一种可行且有效的实现无线传感器节点在空旷环境和障碍物条件下无线信号衰减原理障碍物监控的方法,实现定位与追踪.利用阴影衰落模型建立接收信号强度测量值线性系统模型,并采用SPIN令牌环通信协议收集接收信号强度;创新性地引入最小角回归算法与最小绝对值收缩和选择因子算法(Least absolute shrinkage and selection operator,LASSO),提高了图像重建速度.即在吉洪诺夫正则化与l1正则化算法分析对比前提下,创新性引入改进的最小角回归(Least angle regression,LARS)重建模型与算法,保证重建效果与复杂LASSO算法相似的同时,将重建图像速度提高一个数量级.实测基于16平方米范围内的16个JENNIC 5139节点进行定位与追踪.实测结果与仿真相比虽稍有偏差,但近似符合.这充分表明:吉洪诺夫正则化与l1正则化适用于不同分辨率场景,且都可较好地反映障碍物状况.The emerging technology, radio tomographic imaging （RTI）, uses the attenuation characteristic of wireless signal to locate and trace objects. A linear model with the SPIN communication protocol for received signal strength （RSS） measurements is presented in this paper to get objects0 image in our deployed RTI system. To improve the image reconstruction speed, least absolute shrinkage and selection operator （LASSO） algorithm of compressed sensing field is referred to and compared. Moreover, modified l1-norm regularization is adopted to enhance the resolution of image reconstruction, which is compared with Tikhonov. Based on sixteen JENNIC 5 139 sensor nodes, some experiments have been developed for imaging and tracking the objects inside an area of sixteen square feet. Although there are some differences between simulations and real experiments, the positions of objects can be accurately located from both simulations and real measurements.

关 键 词：无线射频层析成像 重建算法 最小角回归算法 l1最优化 吉洪诺夫正则化 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]