【作者】 王刚刚；

【导师】 王哲龙；

【作者基本信息】 大连理工大学 ， 控制理论与控制工程， 2008， 硕士

【摘要】 目前,随着社会科技的不断进步,我国的医疗服务行业发展迅速,极大的满足了广大人民群众日益增长的医疗消费需求。在当前信息科技不断进步,无线信息传送网络不断完善的同时,医疗卫生领域出现了“远程医疗”这一崭新概念,随之引起了医疗行业的重大变革。远程医疗就是在这种医疗和信息技术不断进步的背景下出现的,是医疗水平和信息技术不断完善的产物。远程医疗的出现为人类提供了更加迅速便捷的医疗服务。据相关资料报道,目前老年人、慢性疾病患者及术后病人的日常监护问题已经成为当前远程医疗服务行业的一个热点问题。本文就远程医疗监护中的人体行为跟踪这一问题进行了一些研究,主要对人体的日常生理状态做实时的监控并对一些异常的步态信息及时报警。同时将人体的一些重要日常行为信息和生理信息做出记录,保存到个人专有数据库中。这样就为将来的医生就诊提供了良好的诊断依据。在上述行为识别监护过程中我们将贝叶斯智能算法引入到行为和步态的识别中,取得了不错的识别效果。本文的人体行为识别过程首先通过无线人体局域网采集人体生理信号,然后对采集来的人体生理信号进行分类处理和判断,将信号离散化。再结合贝叶斯算法达到对一些常见人体生理状态的识别。实验证明,该贝叶斯算法能够有效的进行人体生理状态的识别监控。并且该算法简单可靠,对原始信号数据的要求不高却能达到较高的识别率。此外该算法也可以应用到BSN领域的其他方面。更多还原

【Abstract】 At present, with constant progress of scientist and technology, China’s health care industry has been developed rapidly, which meets the growing demand of people for medical services. Meanwhile, the improvements of information technology and wireless information network caused a major change in the area of medical services. The concept of tele-medicine appears in this background. Tele-medicine can provide even more rapid and convenient medical services through the wireless network technology and the current level of advanced medical care for human beings. It is reported, currently, the custody of the elderly, chronically ill patients and patients after operation has become a major problem in the social health care industry.Some research on the tracking of human behavior of Tele-medicine has been done in this thesis, which includes real-time monitoring and recording of the conduct of human body’s day-to-day state and warning when abnormal gait information is to be found. At the same time, to do a record of day-to-day information and physiological information and the preservation of individual proprietary database will provide a good basis for the diagnosis in the future of the medical treatment. It has made good identify results, which advanced intelligent Bayesian algorithm into the behavior and gait recognition in the above mentioned tele-monitoring process.In this thesis, the process of identification of human behavior includes collecting physiological signals through the wireless local area network firstly, and then classifying the collection of physiological signals, conversing victory signals into discrete ones, at last achieving some common recognition of the state of human behavior combined with a Bayesian algorithm. The experiments have proved that the method using Bayesian algorithm has been greatly improved recognition accuracy of human act. The algorithm is simple and reliable, asking with less original data signal but higher recognition rate. In addition, the method can be applied to other areas of BSN.更多还原