振幅与相位信息相结合的可迁移压缩感知被动式目标定位方法研究

There is a strong need for device-free localization approach with high accuracy and high robustness for behavior monitoring, like rare wild animals monitoring, security protection and etc. However, current approaches face several challenges, including the low reliability in network, the variety in targets, environmental difference and time-varying signal. To cope with these challenges, this project tries to improve localization accuracy by combining amplitude and phase signatures, and to use compressive sensing theory to prolong the life cycle, by reducing energy consumption and improving robustness. To improve the usability, we try to use transfer learning theory to reduce the human efforts in deployment and measurement, which are caused by temporal variation of signal and changes of targets species and environments. We are planning to verify the efficiency and practicability of our solutions in real scenes by monitoring golden monkeys in Qinling Mountains and carrying out related experiments. The scientific essence of this project is to discover two important relationships by using compressive sensing theory, signal processing technique and machine learning theory. One is the inherent relationship between wireless signal and the spatio-temporal features of mobile targets, the other is the relationship among localization accuracy, robustness and the temporal variability of environment. The research findings of this project will provide theoretical support and application reference value for behavior monitoring field in wireless network and have a positive effort on the new generation of human-computer interaction technique.

在珍稀野生动物监测和人员安防等行为监测无线网络应用中，迫切需要高精度、强鲁棒性的被动式目标定位方法，而现有的定位方法却面临着网络可靠性差、目标多样性、环境差异和定位信号时变等挑战。本项目试图利用复信号中振幅和相位信息相结合的思路提升定位精度；引入压缩感知理论降低能耗、提高鲁棒性，解决定位网络的可生存问题；利用迁移学习理论降低因信号时变、目标种类和部署场景变化所引起的巨大人工勘测标定开销，提高定位模型的可用性；项目将通过秦岭金丝猴行为监测以及相关实验验证所提方法在真实场景下有效性与可行性。其科学实质是在被动式目标定位引入压缩感知、信号分析和机器学习理论，寻求无线信号与移动目标时空关系的内在规律，探索定位精度、鲁棒性和环境时变性之间的本质联系，研究成果有望为无线网络行为监测应用提供有价值的理论支撑和应用参考，并对新一代人机交互产生积极影响。

本项目针对传统的被动式定位技术中，仍然面临着商用无线设备中信号扰动大、有效信息获取难，多径环境下定位精度提升难，以及定位网络部署代价和人力成本高等问题的多重挑战，研究了满足多重约束条件下高效的被动式目标感知、定位方法、物联网并发传输等科学问题。本项目无线信号感知中，首次实现了利用低成本商用WiFi，RFID等无线设备的相位和振幅信息，对目标材质识别和成像，大大提升了机场等安防场景中的监测范围。项目提出的基于多径信号空间谱的定位方法，有效利用多径突破了真实环境下定位算法鲁棒性差，精度提升难的问题。项目在无线网络协议及传输研究中，提出的基于环境信息同步算法，极大的降低了网络通信的能耗和带宽开销，很好的解决定位网络部署代价和人力成本高等问题。该项目的多个成果先后发表在Mobicom,ICDCS,Sensys,Inofocom,TON,TMC等高水平国际会议和期刊上，并被MIT,普林斯顿,UIUC等多个国际知名高校引用。部分成果成功的应用在陕西省明长城、秦岭金丝猴野外基地等.相关成果被华商报，今日头条等媒体广泛报道。本项目的研究对物联网和无线传感网产业发展和升级有着重要意义。

内容获取失败，请点击重试