基于柔性传感器老年人动作识别智能监护服装关键技术研究

To monitor the old in a distance and make them free activities in loose environment, some critical technology in it are the focus of our research. The project team have accumulated some experiences on identifying qualified drivers and human activity before this time. On testing of flexible sensors we also have done some work. But some key problems need us making research in detail. Therefore, some research contents for it are listed as follows: 1.Determing which kind of flexible sensors we should choose and which parts on human body we can obtain appropriate biomedical signals when they make activity; Making clear the mechanism of measurement between the biomedical signals and human walk; Establishing the quality relationship between flexible sensor materials and biomedical signals; Determining the relational model between biomedical signals and human walk.2.Determining the recognition method of walk signals for flexible sensor; Optimizing classifier by computation, analysis, improvement; Improving recognition rate by modifying the method of data processing and detection of classifier. Innovation: Presenting a measurement method of human activity by wearable flexible sensors; Establishing Determining the theoretical bases of measuring active with wearable flexible sensors; It is clear and definite that sampling and collecting method of biomedical signals of human activity by wearable flexible sensors..In practice, we take the aged's walk and fall for example, which sampling signals by wearable flexible sensors, to inspect the application method and key technology on detection and identification.It is necessary proving the process right and creating foundation for the practical application on monitoring the aged's daily life.

为使老年人在无拘束、无不适的情况下自由活动，一旦发生意外可以及时得到救助，其中动作信息的采集和识别是关键要解决的问题。对此项目组前期，在人体动作识别方面（如驾驶员握力识别、动作识别）做了一些基础研究。在柔性传感器方面做了初步试验。但一些关键问题还有待进一步研究解决，对此本项目研究内容：1.柔性传感器与人体动作生物信号的选取；生物信号与人体走步间的测量机理；柔性传感材料与生物信号之间定量关系模型的建立；生物信号与人体走步间的对应关系模型。2.柔性传感器走步信号的识别方法；分类器算法比较、选取、改进；提高识别率数据处理和运算方法。创新之处，提出可穿着柔性传感器获取走步信号检测方法；建立可穿着柔性传感器走步动作检测理论依据；柔性传感传感器对人体下肢动作生物信号的拾取方法。实践上，以老年人走步和跌倒为案例，研究可穿着柔性传感器动作检测与识别的具体实施方法和关键技术基础，进行必要的实验验证。

随着我国社会老龄化程度的不断加深，老年人的日常健康监护也越来越引起人们的重视。而老年人的运动失衡，骨质疏松以及骨骼畸变方面的问题较为突出，国内外研究机构都非常重视这些问题的研究。为此，我们基于柔性传感器对老年人动作识别智能监护服装关键技术进行了深入研究。这项研究分为柔性传感器与人体生物信号之间的关系模型；生物信号与人体动作的对应关系模型；柔性组合传感器对老年人日常动作信号的识别原理和方法研究。最后通过实验检验动作识别装置，改进信号处理方法，进而改善和提高分类器识别精度。在此过程中获得了一系列研究成果。