Design of sleep monitoring alarm device based on S3C440X

1 Introduction

The pulse and heartbeat of a normal person are consistent. The pulse frequency is affected by age and gender. It is 120-140 times per minute for infants, 90-100 times per minute for toddlers, 80-90 times per minute for school-age children, and 70-80 times per minute for adults. In addition, exercise and emotional excitement can increase the pulse, while rest and sleep can slow the pulse. When the pulse rate of an adult exceeds 100 beats per minute, it is called tachycardia; when it is less than 60 beats per minute, it is called bradycardia. There are many diseases in clinical practice, especially heart disease, which can cause changes in the pulse. Therefore, measuring the pulse is an indispensable examination item for patients. Traditional Chinese medicine uses pulse palpation as the main method for diagnosing and treating diseases. Diagnosing pulses and treating diseases is an unfathomable knowledge in traditional Chinese medicine. Diagnosing pulses and treating diseases is an extremely valuable experience accumulated by many old Chinese medicine practitioners through their unremitting efforts throughout their lives or even generations. Sleep disorders are the main cause of the formation and development of many diseases. Difficulty falling asleep is the most common phenomenon of insomnia, and most of the reasons are caused by the inability to suppress the continuous thoughts. How to effectively monitor people's sleep state, control "thoughts" during sleep, improve sleep quality, and provide reliable guarantees for the early diagnosis and prediction of sudden diseases deserves special attention. With the development of science, the beating of the pulse can be displayed and recorded very accurately through sensors using microprocessors. By comparing and analyzing the pulse of a normal person and the pulse of the measured object, the physiological condition of the measured object can be quickly and accurately obtained, and an alarm can be issued to relevant personnel in case of danger. This research is based on the cross-penetration of multiple disciplines and is a perfect combination of computer technology, biomedicine and signal processing technology.

2. PVDF film pulse sensor

Although the PVDF film pulse sensor is suitable for the requirements of pulse measurement, it has many disadvantages: 1. It has position correlation. Because it is a pressure sensor, only when the sensor is accurately located at the surface artery where the pulse fluctuation exists can there be a pulsating signal output; 2. The sensor is highly sensitive and will amplify the micro-motion signal, so the random movement of the human body is easy to introduce interference; 3. The sensor shell is made of metal copper, and direct contact with the skin will introduce 50Hz power frequency interference. Therefore, the analog signal output by the sensor cannot correctly reflect the surface pulse fluctuation caused by the artery, and the signal needs to be processed to a certain extent.
 

Figure 1 Schematic diagram of hydraulic capsule

In view of the above shortcomings, a rubber fully enclosed liquid capsule was designed by using the property of "liquid transmission pressure". Figure 1 is a schematic diagram of the hydraulic capsule. The capsule is flat and has a rubber wall thickness of 0.5mm and a cavity thickness of 2mm. The sensor is installed at A in the figure, and the sensor's sensing surface is attached to the upper wall of the capsule, and the lower wall of the capsule is close to the skin. The elasticity of the rubber is used to make the capsule fit the wrist skin, increasing the sensor's sensing area. The sensor can be worn on the wrist by inserting the wristband from both ends B. As long as the hydraulic capsule has a small area in contact with the radial artery of the wrist, the pulse fluctuation will be transmitted from the liquid in the capsule to the sensor's sensing surface, solving the problem of sensor position correlation. Using gasoline engine oil as the pressure transmission medium, the experiment found that: the physical properties of the oil are viscous, and some weak transient fluctuations are not easy to be transmitted, which plays a role in anti-interference and low-pass filtering; the oil has a large specific heat capacity, and its temperature does not change significantly with the outside world, which enhances the wearing comfort. Engine oil is viscous and has a large inertia. It can block random disturbances in the skin background to a certain extent, and also attenuate the transmission of the pulse to a certain extent. Actual measurements have found that the attenuation of "strong" and "weak" in a fluctuation cycle is different. This problem is reflected in the large deviation of the valley shape in the obtained pulse waveform. This problem can be corrected using software. In addition, the hydraulic capsule is made of insulating thermally conductive rubber that is non-irritating to the skin, which isolates the direct contact between the metal shell of the sensor and the skin, and the 50Hz interference is also isolated. Thermal conductivity makes it possible to conduct the surface temperature at the wrist, increasing the measurable physiological signals.

3. Sensor signal processing

The electrical signal output by the pulse sensor is too weak and must be amplified before being output to the analog-to-digital conversion element. The low-pass amplifier is designed based on the maximum number of human pulse beats after exercise, which is 240 times/min. It is composed of IC2A and C04, as shown in Figure 2. The turning frequency is determined by R07, C04, R08 and C05, and the amplification factor is determined by the ratio of R08 and R06.

Figure 2 Low-pass filter diagram

According to the transfer function of the second-order low-pass filter: the gain is: , take 0.707 times the zero-frequency gain to calculate the high-frequency corner frequency: , considering the highest human pulse , the low-frequency characteristics are satisfactory. In order to improve the accuracy of the measurement, the circuit in Figure 3 can be used.

The research object of pulse signal analysis is signal, especially non-stationary and time-varying signal. The signal analysis method can adopt time-frequency analysis. Its main task is to describe how the spectrum content of the signal changes with time, so as to represent the energy and intensity of the signal in time and frequency at the same time. After obtaining this distribution, various signals can be analyzed and processed, the characteristic information contained in the signal can be extracted, and the signal with the desired video distribution characteristics can be obtained by synthesis. Here, the time domain analysis method is adopted.

Figure 3: Secondary amplifier and comparator

The time domain analysis method mainly analyzes the relationship between the height of the pulse wave amplitude and the pulsation. It is currently the most widely used analysis method. In previous studies, most of them used manual reading, that is, using measuring tools to directly read the pulse wave, the height of the gorge, the corresponding time value, the pulse area and other parameters. By analyzing the amplitude and time value of the pulse, we can understand the frequency, rhythm, strength of the pulse, the strength of the pulse and the shape characteristics of the pulse. It extracts some points with obvious physiological significance in the pulse curve as characteristic points for evaluating the pulse wave, which is intuitive and easier for clinical physicians to accept.

Due to the efficient mathematical computing ability and complete functions of Matlab, in this study, the mathematical tools of Matlab were used to directly process the collected signals to obtain the time domain spectra shown in Figures 4 and 5, abandoning the process of manual reading of the graphs and reducing the errors caused by empirical estimation.

Figure 4 Time domain waveform of a healthy person's pulse wave

From the time domain diagram of the pulse wave of a healthy person, it can be seen that the main wave (P wave) and the dicrotic wave (D wave) are obvious, and there is no high-frequency contraction string phenomenon, which is confirmed to be a normal pulse of a healthy person. [page]

From the time domain diagram of hypertensive patients, it can be seen that the main wave peak is flat and wide, the dicrotic wave and the main wave merge into a wide square wave, the descending isthmus is elevated, and the dicrotic wave is flat, which is confirmed to be a string pulse of hypertension, and should be mid-term hypertension.

Figure 5 Pulse wave time domain waveform of a patient with hypertension

The pulse wave time domain graph of insomniacs is between the above two situations, which is easier to judge clinically.

4. Design ideas

In order to solve the problem of being unable to suppress the continuous thoughts, we can let the computer provide appropriate stimulation in a certain way according to the state of the brain to stimulate the source of uncontrolled thinking activities. It is possible that the attention to it is continuously diverted and it is in a suppressed state until it is finally inactivated. The machine-brain interface is such a system that intervenes in the brain's thinking activities through computer monitoring and control. Its function is to interrupt the uncontrolled thinking activities in the transition process from wakefulness to sleep. These functions can be realized by "micro embedded computer systems".

Figure 6: Machine-brain interface structure diagram

The machine-brain interface is a system that intervenes in brain thinking activities through computer monitoring and control. Its function is to interrupt the uncontrolled thinking activities in the transition from wakefulness to sleep. In order to achieve this function, an interrupt source is first required. Here, a stimulator is used to generate sound stimulation as an interrupt source; after the stimulation is generated, the user's response to the stimulation is converted into an electrical signal through a sensor; the weak signal from the sensor is amplified and filtered and sent to the embedded computer system for various processing. In the event of sudden illness (such as cerebral thrombosis), the embedded computer system can send out a "sound" alarm or a "voice" telephone alarm through remote control to notify family members and emergency centers in time. The overall structural block diagram of the machine-brain interface system is shown in Figure 1.

5. Hardware Design

S3C440X is a 32-bit high-speed processor based on the RISC ARM7TDMI core launched by SAMSUNG. Its operating voltage is 3.3V, and the operating voltage of the ARM7TDMI core is only 2.5V. Such a low-power chip is more conducive to the development of wearable products.

The hardware connection of S3C44B0X to realize sleep monitoring is shown in Figure 7.

Figure 7 Sleep detection hardware connection diagram

      The circuits for sound alarm and telephone voice alarm are relatively common and will not be described in detail here.

6. Software Design

Linux is a very popular operating system, which is compatible with UNIX system and open source. It was originally designed as a desktop system, but is now widely used in the server field. The greater impact is that it is gradually being applied to embedded systems. uClinux was born in this atmosphere. It is Micro-Control-Linux, which literally means "Linux system designed for the micro-control field".
 
The specific system transplantation will not be introduced in detail here. Here is the software control block diagram of the system, as shown in Figure 8:
 

Figure 8 Software control block diagram

7. Conclusion

The research on sleep monitoring alarm system based on thin film pulse sensor, based on the study of sleep transition stages, emphasizes the function of timely realizing sound alarm or telephone voice alarm if sudden heart and brain diseases occur during sleep.

Previous sleep research was mainly based on monitoring. The innovation of this project lies in the following two aspects.

1. The dual sensor structure is used to measure and monitor the electrocardiogram (ECG) of the sleeper to test the cognitive response of the subject, so as to wake up the light sleeper and conduct cognitive testing on the seriously ill sleeper.

2. Sound alarm and telephone voice alarm. Based on the research on electrocardioencephalogram (ECG) in the sleep state, if the sleeper has sudden heart and brain diseases, the EEG will be displayed in time and the wireless alarm will be triggered through the embedded processor to give a timely alarm.

references

[1] Wang Ruiying et al. Sleep monitoring control system [J]. Chinese Journal of Clinical Rehabilitation 2006; 10 (13): 119-121
[2] Yu Mengsun Zhang Hongjin. A new model of sleep medical monitoring [J]. China Medical Equipment Information, 2003, 9 (3): 4-6
[3] Zhang Dabo. Embedded system principle, design and application [M]. Beijing: Machinery Industry Press 2005: 448-449
[4] Zhang Yong. Sleep monitoring system [G]. Master of Engineering Thesis of South China University of Technology 2005: 27-28
[5] Dou Zhenzhong, Song Peng, Li Kai. New development and challenges of embedded system design [J]. Microcontroller and Embedded System Application, 2004, (12): 15-18
[6] Lin Guimao. A new method for monitoring sleep transition process [J]. China Medical Equipment Information, 2005, 11 (3): 29-32
[7] Jiang Xun, Hu Youlun, Zhang Wei. Hardware design of pulse measurement instrument [J]. Medical Equipment Information, 2005, 20 (7): 1-3
[8] Yuan Qiming. Sleep monitoring system [J]. Medical and Healthcare Devices, 2006, 3: 1-2