Research Of Discriminant Technology For Human Body Physiological State Based On Multi-sensor

In recent years, more and more wearable products used to monitor the human health are constantly entering people’s lives with the rapid development of bio-sensor technology and wearable technology. Acquiring and recording the physiological health data such as pulse, respiration and body temperature have become realistic and feasible. However, physiological data in a continuous long period of time is too large. A large amount of data is not conducive to analyze, observe and extract valuable information of the human body. Therefore, the purpose of this study is to achieve a discriminant technology. It can extract the concise but effective changes on human physiological state from the continuous and large amounts of human physiological data.The body’s physiological state can be broadly divided into two categories:one is the normal state, which the body is in a resting state; another is the event state, which people may experience activities, external stimulation or mood changes. The discrimination method of body’s physiological state can be carried out as a dichotomous classification. It needs three signals:pulse, respiration and body temperature, each signal has its own discriminant mechanism. The classification result is a visualization figure of levels distribution. Users can focus on the physiological data that they are interested in according to their state levels.In this thesis, the discriminant mechanism of respiratory signal and temperature signal is to set the threshold. By judging whether the respiratory rate and body temperature could exceed the normal threshold range, the physiological state of the body was dichotomous. As pulse waveform often contains a lot of high-frequency interference and baseline drift, we use the method based on discrete wavelet transform to get rid of them. Then, we extract the pulse period and height of systolic peak from the time domain as the input feature vectors of support vector machine (SVM). Through a binary classification model built by the method of supervised learning, the physiological state is judged as normal state or event state. Final, we take three experiments:movement, sleep and drink. The statistical analysis and evaluation results show that the classification performance of SVM is excellent. After the fusion of three signals, visualization software can show users the state levels with the distribution of time, which provides an overall and concise perspective to observe changes in physiological state.