利用多维链路信息及环境无线信号的被动定位与状态识别方法研究

Device-free wireless localization and status recognition is a technique that could realize localization and status recognition of a target by analyzing the shadowing effect of the target on wireless signals, while removing the requirement of equipping the target with any device. It has become a new field of wireless localization research, and could provide fundamental data for pervasive computing and smart environment applications. Existing work realizes device-free wireless localization and status recognition by analyzing the variation of received signal strength incurred by the presence of a target. This proposal investigates the new thought of achieving device-free wireless localization and status recognition by utilizing multi-dimensional wireless link information and ambient wireless signals, e.g., FM, GSM, and WiFi, etc.. Firstly, the method of achieving high performance device-free localization and status recognition by analyzing multi-dimensional wireless link information, e.g., received signal strength, time of flight, and channel state information, is presented. We build 3D saddle surface model to represent the relationship between link information and target’s location, propose Bayesian graphical algorithm to realize location estimation, and realize status recognition based on multi-dimensional feature information. Furthermore, the method of achieving device-free localization and status recognition by acquiring and analyzing the multi-path effect of the received ambient wireless signals is presented. We utilize beam forming technique to acquire multi-path information, and realize location and status recognition with a novel differential dynamic time warping technique. The above work would provide novel methods for device-free wireless localization and status recognition technique, and provide theoretical basis for location dependent applications.

被动无线定位与状态识别是一种利用目标对无线电波传播的遮蔽效应实现位置估计与状态识别的新方法。与传统方法相比，该方法无需目标携带任何设备参与计算，已成为无线定位研究的新领域，可为普适计算、智能环境等应用提供基础数据。已有工作利用目标对无线链路信号强度的影响实现定位与状态识别，本研究探索了利用多维无线链路信息以及自然空间环境无线信号（FM、GSM、WiFi等）实现被动定位与状态识别的新思路。首先，提出利用接收信号强度、电波传播时间及信道状态等多维链路信息的高精度位置估计与状态识别方法，建立三维马鞍面模型描述链路与目标位置关系，提出贝叶斯图形算法实现位置估计，利用多维特征进行状态识别。其次，提出利用环境无线信号多径传播特征的位置估计与状态识别方法，采用波束成形技术捕获多径特征，利用差分动态时间规整方法进行位置状态估计。研究成果将形成新的被动定位与状态识别方法，为其在相关领域的应用提供理论基础。

本课题探索了利用多维无线链路信息以及自然空间环境无线信号（FM、GSM、WiFi等）实现被动定位与状态识别的新思路，设计了相关的算法、方法,并通过理论分析与实验验证的方法充分了验证了系统的可行性。在自然科学基金的资助下，课题组发表学术论文9篇，其中，SCI期刊长文论文7篇，IEEE汇刊6篇。所提出的基于小波特征的高性能被动定位与状态识别方法发表在IEEE Trans. Vehicle Technology (IF 4.006中科院二区，JCR一区)，提出的节能快速的被动无线定位识别系统发表在IEEE Trans. Industrial Informatics (IF 6.764 中科院一区，JCR一区)，提出的基于二维特征的被动无线定位与状态识别方法发表在IEEE Trans. Vehicle Technology (IF 4.006中科院二区，JCR一区)，提出的被动定位与状态识别无线网络内数据的高效传输方法发表在IEEE Trans. Vehicle Technology (IF 4.006中科院二区，JCR一区)。同时，课题组获批国家发明专利4项。课题成果推动了被动无线定位与状态识别技术的发展，为其在智能空间、智能家居等场景的应用提供了方法支撑。