Development for automatic diagnosis system for health condition, using entropy of behavior information

在第一年，我开发了测量系统和数据采集和分析软件，测量系统由红外传感器、有线和无线网络系统以及蓝牙无线通信设备组成。并利用加速度传感器和蓝牙无线通信设备开发了移动式传感器系统。在传统的测量系统中，测量系统存在体积大、能耗大的问题。在所开发的系统中，为了便于携带，减小了设备的厚度。改进后，延长了测量时间。此外，我还可以在互联网上获取该课题的数据，因为我在WWW系统上开发了数据采集软件。此外，我还开发了一种从长期数据中获取熵的算法。在第二年，我验证了所开发系统的实用性和有效性。在这一年里，我获得了日常生活中行为的长期数据。另外，对长期数据进行了分析，首先采用了传统的线性方法进行比较。然而，我发现了数据的异常点，我找不到正常数据和异常数据之间的显著差异。其次，利用长时间数据的信息熵来获取指标。我发现这些指标与正常状态和异常状态之间有显著差异。这一结果与问卷调查结果相吻合。结果表明，该建议技术是有效的。

In the first year, I developed measurement systems and software for data acquisition and analysis.The measurement system consists of infrared sensors, wired and wireless network systems, and Bluetooth wireless communication device. I also developed the mobile type sensor system using accelerometer and Bluetooth wireless communication device. In the traditional system, the measurement system has problem of the size and energy consumption. In the developed system, I reduce the thickness of the device for the portability. As the result of improvement, the time for measurement is extended. In addition, I can acquire the data of the subject on the Internet, because I developed the data acquisition software on the WWW system. Furthermore, I developed an algorithm to acquire the entropy from long term data.In the second year, I verified utility and the validity of the developed system. In this year I acquired the long term data of the behavior in the daily life. In addition, I analyze the long term data First I adopt the traditional linear method for comparison. However I found the unusual points of data, I could not found the significant difference between usual data and unusual data. Second, I acquired the indexes using entropy of long time data. I found the significant difference between the indexes from usual and unusual state. This result was corresponding to the result of questionnaire survey. As the result, the effectiveness of the proposal technique was suggested.