车辆三维无线层析成像的识别与定位方法研究

The state-of-the-art localization technique based on radio tomographic image holds trouble to identify the type of the object and is not capable of locating the high-speed object with big size. In order to solve these problems, relying on theory of radio tomographic image, an identification and localization method of three dimensional radio tomographic image for high-seed vehicle is proposed by the research work on self-localization method of three dimensional senor network with extraordinary topology and the rapid construction method of three dimensional radio tomographic images. Firstly, an accurate received signal strength measurement method based on the weight of fade level, and a selection mechanism of the reference plane to anchor node in the wireless sensor network with extraordinary deployment are introduced to achieve accurate self-localization of the three dimensional sensor nodes. Secondly, a topology construction technique of three dimensional radio tomographic image is employed based on analyzing qualitative relation among the sampling rate of sensor node, the speed of vehicle, the number of sensor node and the size of the three dimensional voxel. Thirdly, the rapid construction method based on multi data frames to three dimensional radio tomographic image is studied by constructing the weight model of the three dimensional radio tomographic image for vehicle. Finally, the experimental verification platform is established for the aforementioned methods and models. Achievements in the identification and localization method of the three dimensional radio tomographic image for vehicle and theoretical model will provide new approaches and effective solutions in the applications of battle perception, national defense and industry.

针对现有无线层析成像定位方法不能有效识别目标类型、不能适应定位快速移动的大体积目标等问题，本项目基于无线层析成像理论，通过三维异构传感器节点自定位算法和多数据帧三维层析图像快速构建算法的研究，提出一种车辆三维层析成像的识别与定位方法。首先，提出基于衰减尺度权重的通信链路接收信号强度测量方法和三维异构传感器网络锚节点参考平面的选择方法，解决车辆三维层析成像节点的自定位问题；其次，在分析节点采样频率、车辆速度、节点数量与三维网格大小定性关系的基础上，提出车辆三维层析成像的网络拓扑构建方法；再次，建立车辆三维层析成像的权重模型，研究多数据帧车辆三维层析图像的快速构建方法，并通过三维层析图像的分析，实现车辆的识别、定位及测速；最后，研究建立系统的原理试验验证装置。可望在车辆三维层析成像的识别与定位关键技术、理论模型等方面取得创新性成果，将为相关技术在战场感知、国防及公安领域的应用提供重要理论依据。

利用战场环境中无线电磁波信号的扰动来识别、定位、跟踪丛林环境中车辆等目标列为我国军方亟需的战场监测技术之一。利用三维无线层析成像（RTI）定位技术可实现战场车辆的探测、识别与定位，可在全天候环境下以低成本方式，为军事情报部门及时提供可靠的敌方军事部署及目标动态，确保我军掌握战场主动权。在此需求背景之下，本项目提出车辆三维无线层析成像的识别与定位方法研究，着重研究了以下四方面的内容：.（1）.车辆三维RTI节点自定位算法，提出基于概率密度的RTI节点自定位方法和基于四面体体积优选锚节点的三维无线节点自定位方法，将现有RTI系统传感器节点坐标获取方式由人工测量提升至自动测算，拓展了该定位技术紧急部署/快速定位的应用场景；.（2）.静态车辆三维无线层析成像定位算法，提出一种基于链路投影网格法的车辆三维静态定位方法，通过基于频率分布直方图和核距离的无线通信链路筛选方法研究、基于信息熵的通信链路权重模型构建、基于空间协方差模型的网格权重矫正，与传统方法相比（仅定位人员），实现了大体积静态车辆的三维无线层析成像定位及识别；.（3）.动态车辆三维无线层析成像定位算法。提出一种基于卡尔曼滤波的车辆三维动态定位及测速方法，通过基于长短窗直方图和核距离的链路筛选方法研究、基于卡尔曼滤波的体素权重估计方法研究、车辆三维动态定位图像重建和基于先验估计的车辆测速方法研究，实现了动态车辆的三维无线层析成像定位、识别及测速；.（4）.车辆三维RTI识别与定位方法的试验验证装置。研究设计了适于车辆三维RTI的无线传感器节点间多信道切换令牌环通信协议；汇聚节点多信道切换令牌环接收通信协议及车辆三维RTI定位系统网络拓扑结构。.上述研究成果可推广至国防应用中大体积非合作目标的探测与识别。