基于周边智能环境的人体运动行为感知理论与方法

Human activity sensing is the realization foundation of "Human-centered" Ambient Assisted Living (AAL), and has an important application value in the various fields, such as human-computer interaction, smart environment and security surveillance. Motion recognition is important basis of activity understanding. The goal of our research is to develop a distributed compressive sensing scheme for human activity in an ambient intelligence environment.. Inspired by the compressive sampling theory and the reference structure tomography, a fiber sensor network system for human tracking is implemented, and the boolean compressed fiber-optic sensing model is put forward, which transforms the sampling task in pressure fields from high-dimensional source space to low-dimensional measurement space. With optimizing the deployment of fiber sensor networks from the number of fiber sensors required, the dynamic uniquely decipherable code scheme is employed to guide the fiber sensor deployment for target localization. Infrared radiation changes (IRCs) induced by human activity can provide important information about activity patterns. Based on the compressive sensing paradigm, we propose a compressive infrared sensing approach to fully capture the IRC compressively. In particular, the hierarchical compressive infrared sensing will be investigated for achieving scalable visibility coverage and fulfilling multi-level sensing requirements.. The research has an important academic value and practical significance for the promotion of extension and application of the compressive sensing theory and reference structure tomography and pushing the development of the AAL technology.

人体运动行为感知是实现“以人为中心”环境辅助生活的基础。人体运动行为在周边环境中产生的足迹和辐射能量变化信息是行为分析的重要线索。本项目以该线索为感知目标，将压缩传感理论与几何参考结构层析成像原理相结合，建立基于压缩传感的光纤感知模型和红外感知模型，探索适合分布式网络化实现的多层次运动行为协作感知模式，将目标空间的高维运动信息投影为测量空间的低维数据流。在此基础上，利用低维测量空间内多线索运动信息的抽取和融合，实现多层次人体运动行为分析。本项目解决了分布式无线传感器网络对计算通信资源受限的困境，这对促进压缩感知理论的应用，推动周边智能环境支配的人体运动行为感知机制的发展，具有重要的学术价值和实际意义。

项目以人体运动产生的足迹和辐射能量变化为感知目标，以光纤压力传感技术和热释电红外运动传感技术为感知手段，对压缩传感支配的几何参考结构层析成像进行了探索，分别从压缩传感模型理论和实现方法两个层面，重点研究了压缩光纤运动跟踪技术和压缩红外运动识别技术，突破了传感效率存在的技术瓶颈，形成了适合分布式网络化实现的多层次运动行为协作感知模式。. 此方面的研究，无论是对人体运动行为感知自身的突破发展，还是对压缩跟踪和压缩识别理论的应用拓展，都具有十分重要的学术意义，在智能环境、人机交互、无线传感器网络等领域有广泛的应用价值。