基于可穿戴技术的信息处理与数据融合算法研究

发布时间：2018-02-02 22:26

本文关键词： 可穿戴脉搏采集姿态识别数据融合实时监测信息处理　出处：《北京服装学院》2015年硕士论文　论文类型：学位论文

【摘要】：随着电子信息行业的迅猛发展，智能可穿戴技术以其信息化、智能化、易携带等优势逐步走入人们的生活。无论是在医疗保健行业、还是在野外救援方面均有非常重要的现实价值。但是，目前国内大部分可穿戴设备仍处于基础研究阶段，一些成形的产品技术含量不够，并非用户刚需，产品价格也虚高。对此，本课题基于可穿戴技术的信息处理与数据融合算法，设计了一款完整的具有环境参数及人体生理参数监测的可穿戴信息监测系统。该系统结合无线传感器网络技术与服装艺术设计的优点，打破了传统服装设计的局限性。同时采用无线传感器网络技术全面、高效、经济的布置监测节点，可将人们从线路复杂而受限制的有线设备中解脱出来，使佩戴者具有更高的活动程度和灵活性。本文运用VS2010、.NET、SQL2008、IAR等软件平台等对系统的服装体系架构、心率采集模块、上位机在线信息监测等进行了软件和硬件方面的设计与信息处理；通过对加速度传感器模块三轴数据的分析，提出了不同的人体姿态判别算法。采用基于单一加速度传感器的经验值摔倒算法实现了人体摔倒的基本识别；针对猛然弯腰等运动所引起的误判，提出了基于二次判断的摔倒检测算法；同时结合数据融合理论提出了基于SVM分类算法的人体姿态识别算法，，从而为分类问题提供了良好的泛化能力。为降低噪声对无线信号的干扰，本文采用了卡尔曼滤波算法对人体的不同姿态进行了滤波处理，提高了判别的准确度。本文融合了信息管理平台的优点同时引入硬件控制平台和串口通信技术对在线信息监测系统进行了设计，真正实现了对各无线信息的实时远程监控和用户管理功能，达到了很好的预期效果。智能可穿戴设备作为智能终端产业下的一个热点已被广泛接受，可穿戴设备必将带动行业的发展，开辟更加广阔的市场。
[Abstract]:With the rapid development of electronic information industry, intelligent wearable technology has gradually entered people's life with its advantages of information, intelligence, easy to carry, whether in the medical care industry. However, most of the wearable devices in China are still in the basic research stage at present, some of the formed products are not enough technology content, not the rigid needs of users. The price of the product is also high. For this, this topic based on wearable technology information processing and data fusion algorithm. A wearable information monitoring system with environmental parameters and physiological parameters of human body is designed, which combines the advantages of wireless sensor network technology and clothing art design. At the same time, the wireless sensor network technology is used to lay out the monitoring nodes comprehensively, efficiently and economically, which can free people from the wired equipment with complex and restricted lines. To give the wearer a higher degree of activity and flexibility. In this paper, the clothing system architecture, heart rate acquisition module and other software platforms such as SQL 2008 IAR and VS2010 / Net are used. The software and hardware design and information processing are carried out on the on-line information monitoring of the host computer. Based on the analysis of triaxial data of acceleration sensor module, different human attitude discrimination algorithms are proposed, and the basic recognition of human fall is realized by using the experience value fall algorithm based on a single acceleration sensor. In view of the misjudgment caused by sudden bend motion, a fall detection algorithm based on second judgment is proposed. At the same time, combining with the data fusion theory, a human body attitude recognition algorithm based on SVM classification algorithm is proposed, which provides a good generalization ability for classification problems and reduces the interference of noise to wireless signals. In this paper, Kalman filter algorithm is used to filter human body's different postures, which improves the accuracy of discrimination. This paper combines the advantages of the information management platform and introduces the hardware control platform and serial communication technology to design the online information monitoring system. The function of real-time remote monitoring and user management of wireless information has been realized, and the expected effect has been achieved. As a hot spot under the intelligent terminal industry, intelligent wearable devices have been widely accepted. Wearable devices will certainly drive the development of the industry and open up a wider market.
【学位授予单位】：北京服装学院
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP274.2;TP202

【参考文献】