Detecting Physiological Parameters Based On Micro-movement Sensitive Mattress Sleep Monitoring System

Sleep is an important physiological state and the quality of it is quite significant for human health. Medical researches in recent twenty years have proved that, many substantial diseases of human are often related to SAS in sleep, such as hypertension, coronary artery disease (CAD), cardiac arrhythmia (CA), diabetes mellitus (DM), cardiac-cerebral vascular accident, psyche affection, and so on. Therefore, it is important to develop sleep monitoring technique.At present, the traditional sleep monitoring is polysomnography(PSG) which contains electroencephalogram, electrooculogram, eletromyogram, and so on. Then computer estimates sleep stages according to international standard. Also PSG is gold standard in sleep monitoring, it needs to affix at least 10 electrodes on patients'bodies, which make great physical and mental stress on them.The Micro-movement Sensitive Mattress Sleep Monitoring System (MSMSMS) is an unconstrained sleep monitoring system that can continuously measure body movement, respiratory wave, heart rate without sticking any electrode on patients' bodies, and it can also diagnose sleep apnea syndrome(SAS), obtain the whole night sleep structure. As MSMSMS is of benefit to extend sleep monitoring, develop sleep and physiological research, it has great useable potential. This system can be applied to daily and long time sleep monitoring, such as monitoring the sleep quality of special workers(airman, engine driver, athlete, and so on).By the improvement of hardware, the new system has unparalleled advantages compared with the previous. The work of this paper is mainly based on the new system to pre-process the hardware outputs, and then extract biological parameters, including leave bed, body movement, respiration rate, heart rate. All of them can provide judgment for sleep stages and sleep-breath events. The main tasks of this article are:(1) Establishing pre-processing methods based on the new system, and then getting ballistocardiography, respiration wave, leg impact. Respiration wave and leg impact are easy to acquire compared with ballistocardiography, as the special character of the bed system. This paper used two methods to process the original signals which are wavelet transform and digital filter, the results are very well and the character can be seen clearly.(2) Detecting physiological parameters, including leave bed, body movement, respiration rate and heart rate. The first three parameters have been used in previous system. However, the algorithms of them are relatively complex and make the software run slowly. This paper proposes relatively simple algorithms to increase its run speed and the results are also very well. The heart rate from the bed system is first used and this makes unconstrained sleep monitoring come true. However, the sleeping position and weight of the users can affect the bed signal seriously, so heart rate detecting is hard and also is this paper's important work. This paper used the window threshold method, multi-channel information fusion and other methods to detect physiological information. By comparison of large amounts data, we confirm that these extraction algorithms are easy and useful, the results are satisfying.(3) Extracting sleep structure of SAHS using heart rate from the bed system. This paper calculated the fractal exponents of different sleep stage using detrended fluctuation analysis(DFA), and the results show that the fractal exponents have significant differences in different sleep stage for SAHS patients, and they are also different between healthy subjects and SAHS patients. All of these supply references to distinguish sleep stages of SAHS patients used heart rate and diagnose SAHS as well.