Design of Low Power Pulse Oximeter for Early Detection of Hypoxemia

Transcription

1 2016 International Conference on Micro-Electronics and Telecommunication Engineering, Design of Low Power Pulse Oximeter for Early Detection of Hypoxemia Komal Kashish, Matangi Priya, Piyush Yadav Department of Electronics and Communication Engineering G.L Bajaj Institute of Technology and Management Greater Noida, India Abstract This paper involves designing a low power yet effective pulse oximetry system for detection of Hypoxemia especially for patients from developing and under-developed nations where the cost of any medical device is a primary concern. For this non-invasive technique to draw results from the behavior of light absorption pattern shown by blood, we decided on using a simple red LED and an infra-red LED, with wavelength of 660nm and 940nm respectively. We used a receiver module on the other side to collect the refracted light. As predicted, the graph of the refracted light intensity mimicked the heart rate pattern of an individual. This proved to be helpful in drawing results such as heart rate and oxygen saturation level in the blood. Keywords Blood Oxygen Saturation Level; Hypoxemia; Heart Rate; Light Absorption Pattern; Pulse Oximeter; Refracted Light I. INTRODUCTION Despite the catapulted demands of fitness bands in developed nations to measure human body parameters such as heart rate, calories level etc., using these devices remain rather impractical because of their market price when the general population of developing countries such as India is considered. Keeping in mind the required trade-off between accuracy and price of such devices that needs to be done, we thought of designing a system which uses simple electronics and optics principles to give out similar accuracy such as the professionally used as well as the commercial devices. Widespread use of such gadgets can help in minimizing the problem of non-uniform health monitoring in the world and the apparatus acting as an indicator of early signs of health conditions, such as Hypoxemia [6], which could have otherwise gone unnoticed due to the unawareness towards regular health check-ups. The solution proposed here is using optical methods to derive human body parameters which proved to be both, cost effective as well as accurate. II. BACKGROUND INFORMATION ON HYPOXEMIA Hypoxemia refers to extremely low oxygen levels in blood, specifically arterial blood [2]. A broader term referring to oxygen deficiency is Hypoxia, which is abnormally low level of oxygen in other parts of body as well and has multiple impacts on the vascular system. [1]. Hypoxemia is a category of Hypoxia, known as Hypoxemic hypoxia, which is caused when blood s oxygen carrying capacity is reduced, such as in the case of Carbon Monoxide poisoning [4]. Hypoxemia is caused by insufficient oxygen content in arterial blood which may be due to any one of the five known mechanisms: ventilation-perfusion mismatch, diffusion impairment, alveolar hypoventilation, right to left shunt and diffusion-perfusion abnormality. A severe case of Hypoxemia is called Anoxemia. Some of the common symptoms related to Hypoxemia are coughing, accelerated heart rate, delirious state of mind, sweating etc. An early diagnosis of these symptoms can help in combating the deleterious effect of Hypoxemia. III. WORKING PRINCIPLE The basic principle behind a pulse oximeter is simple. We used LEDs with wavelengths of 660nm and 940nm. We chose these two LEDs because at these particular wavelengths only the oxygenated and deoxygenated blood play the most important role in light absorption when compared with absorption at other wavelength values where the absorption coefficient of other layers present in the body, tissues and muscles etc. is high [5]. Also at these two wavelengths, the absorption levels of deoxygenated hemoglobin and oxygenated hemoglobin of blood are not similar and can be easily distinguished apart by the values of extinction coefficient, which in molecular biology and chemistry is a measure of how strongly a substance absorbs light at any given wavelength. Fig 1. Absorption Spectrum /16 $ IEEE DOI /ICMETE

2 For example, the emission from LED with wavelength in the 660nm region when projected on the upper surface of the finger will undergo higher absorption by the deoxygenated hemoglobin as compared to oxygenated hemoglobin. Similarly, the absorption of wavelength in the infrared region is much higher in oxygenated hemoglobin than in deoxygenated hemoglobin. This feature of varying power of absorption of blood at these two considered wavelengths correspond to a sinusoidal pattern on an oscilloscope which comes out similar to a heart signal of an individual. This occurs because when the arteries are pumped with blood they expand in size, therefore increasing the absorption of light [9]. Thus, the factor of saturation level of hemoglobin can be calculated as a measure of absorption levels at the two wavelengths. The design of oximeter based on this behavior of blood can be of two types: transmission and reflectance type [10] [11]. In reflectance type oximeter, both the transmitter and receiver are on the same side. The light shone on the area of the body passes through skin and tissues but some part of it is reflected back from the bones and skin. This reflected part is collected by the receiver. In transmission type oximeter, the transmitter and receiver are on different sides, i.e., light from the LEDs is shone on the area selected for oximetry calculations and it reaches the receiver after passing layers of tissues and vessels. Although lower in accuracy than reflectance type oximeter, transmission type oximeter are easier to use as well as are capable of operating at lower light intensity. Hence, because of its low power feature we will use transmission type oximeter model in our device. Third constraint of the system was to minimize the external noise in the system as much as possible. The photodiode used in the study is a highly sensitive component and thus needs to be isolated from external unwanted wavelengths for optimal results. Fourth requirement was to ensure that the size and weight of the designed apparatus remains as less as possible because a bulky system would be inconvenient to use on a daily basis and thus the device would lose its purpose. Last and one of the most important system requirement was that the entire gadget should be inexpensive in installation as well as in operation. V. OUTLINE OF THE SYSTEM Projected on the selected body part, the first part of the unit, transmitter, is built up from a red LED and an infrared LED. On the other side is the receiver unit consisting of a highly accurate photodiode which is then connected to a filter which filters out the unwanted frequencies that the photodiode usually acquires and is capable of functioning at extremely low frequencies, even less than 10 Hz. The working of LEDs and the analog output of the receiver both are connected to a low cost microcontroller, an Arduino ProMini 3.3V in our case, having an inbuilt analog to digital converter for convenience. The values from the Arduino microcontroller can either be sent to an Android App designed to communicate through Bluetooth or to a Liquid Crystal Display (LCD) screen attached with the apparatus. We decided on the former because of the size reduction that occurs by using our Android phones as an alternative to the bulky LCD panel. Also, it restricts the cost by avoiding another hardware addition to the system. Fig 2. Transmission versus Reflactance Type This observation of light absorption and heart signal having similar patterns lead to the conclusion that the human body parameters such as heart rate and oxygen saturation level in blood can also be derived from the optical absorption pattern. IV. SYSTEM REQUIREMENTS Several system constraints are considered when designing the pulse oximeter [7]. First of them was to create a system which can operate at severely low frequencies. This is so because the blood flow in the body occurs at frequencies which are even less than 10 Hz. This constraint called for the use of components designed to work at low frequencies. Second requirement of the system would be for it to function on low voltage sources. Our proposed system is to work on a dc source of mere 9V. Fig 3. Block Diagram of the System

3 The probe placement of the device is a major factor in deriving accurate results from the patient. A probe maybe of finger, ear, toe, sole or palm type [17]. Although a general conclusion cannot be drawn regarding the best region to use the probe, the ease of use in the finger type probe is the highest. Therefore, we opted for a finger type probe design for our device. The probe should be used in such a way that it s well positioned and not too tight on the finger, or else the readings will be disturbed. Other factors which deviate the oximeter readings are, nail varnish, bright external light, movement of the patient etc. Presence of additional pigments on the finger such as nail paints or henna blocks the signal coming from the probe and increase the absorption of the light. Another interfering factor to affect the probe readings would be the surrounding lights, whether sunlight or artificial. If too much of this light is allowed to enter the device, the readings are sure to come out erratic and inaccurate. An obvious factor affecting the oximeter readings is the movement of the patient wearing the probe. It is to be made sure that the patient is completely stable before taking the readings. VI. FINAL SYSTEM AND WORKING The transmitter block has two LEDs, as mentioned above. These are connected to the digital pins of the Arduino. The microcontroller turns the two LEDs on and off at separate times. The two LEDs should never be turned on together at the same time. Resistors are used for intensity control of the LEDs. Since the circuit for the transmitter was simple, to restrict the cost we opted for soldering it directly on a zero PCB board after testing instead of designing a PCB. On the other side of the oximeter is a receiver board. To collect the refracted light that is obtained on the other side of the finger, we required a photodiode which could not only detect the signal but also provide an in-system amplification to reduce the space by avoiding the need of another amplification unit on the board. Thus, we opted for OPT101. OPT101 has an on chip photo detector and a transimpedence amplifier. Specifications- Internal Resistance 1M Bandwidth 14 khz Open loop Gain 90 db Offset Voltage 0.7 V Min. Operating Voltage 2.4 V Following the OPT101 chip in the circuit had to be a band pass filter as OPT101 brings in a lot of undesired frequencies in its output. The filter we needed to design was required to have a bandwidth between Hz. Our first choice for the Op-Amp IC was LM358 as it was readily available to us in the lab. However on testing, the performance of the IC turned out to be quite poor. Hence, as our next choice we implemented the band pass filter with TL- 072, a dual input Op-Amp IC which had better performance in the low frequency spectrum. Fig 5. Breadboard Testing of the Circuit We implemented a two stage amplifier circuit to reduce the noise as much as possible and to smoothen the output and used a diode after the receiver circuit to restrict the negative values in the graph to enter the microcontroller as Arduino reads these analog values as 0 Once testing was done, we needed to implement the entire receiver circuit on a board which was done by designing a PCB. We chose ExpressPCB, an open source PCB designing tool to design the circuit. Fig 4. OPT101 Chip Fig 6. Complete Receiver Layout

4 to be greater than maximum value, the maximum value was set to the current sampled value. Same occurred if the current value was smaller than minimum value. The current value was assigned to Vmin. This process occurred every time the interrupt occurred and we had a new current value. After a few interrupts, we got a Vmin and Vmax value for both red and infrared LED [8]. To further calculate oxygen saturation level of blood by using these obtained maximum and minimum voltage level, the following equations were used. Fig 7. Oscilloscope Graph The entire circuit of the band pass filter was simulated on Cadence Simulation Software by importing the library of IC TL-072. The simulated and the recorded result on the oscilloscope had a very less margin of difference. Thus, it was concluded that the circuit was working the way we wanted. To increase the efficiency of the receiver, we installed the transmitter and receiver circuit inside a closed black enclosure so as to limit the extent of the natural light entering the system. Fig 8. PCB Layout of the Reeciver Once the analog value from the oximeter reaches the microcontroller, the program developed in Arduino IDE (integrated development environment) calculated the oxygen saturation level and the heart rate. To calculate these parameters, we sampled the incoming analog signal every 0.05 sec. Since the signal is periodical, almost sinusoidal, it would always cross a decided value. We concluded that the signal has crossed this set point by comparing two consecutive samples of the signal. If the current sample is greater than the previous sample and is at the set point, then it can be treated as a starting point of one wave and ends when another such sample occurs where it is at the set point and is greater than the last observed sample. To calculate the heart rate, we needed to calculate the total sampled time between these two points, the starting and the end. To calculate the oxygen saturation level, maximum and minimum values corresponding to both red and infrared led were calculated. This was done by assuming two values, Vmax and Vmin. The voltage obtained from the receiver was considered and the current sample was compared with the maximum and minimum value. If the current value turned out Equation no. 2 uses the ratio X to calculate percentage of oxygen saturation in blood (denoted by Sp02). It is a standard model of calculating SpO2 which is commonly used in context of medical devices. Once these two parameters were obtained, the data was sent using the Bluetooth module, HC-06. We decided on using Bluetooth communication because of its heavy availability in the regular phones. Also, it had the advantage of working offline, without the need of the user to always be connected to a server. An alternative to Bluetooth communication could be use of Zigbee systems [12]. Our Android application was developed on MIT Beta AppInventor, an open platform for developing Android applications using simple drag and drop method, eliminating the need of learning complex text-based programming languages for application development. The main advantage of using an Android app using AppInventor, besides the ease in developing the algorithm, are the simple requirements that a phone needs to have to run this application. MIT AppInventor supports all Android systems 2.3 ( Gingerbread ) and higher, and runs on phones which have an SD card, virtual or physical. Fig 9. Our Prototype (1) (2)

5 To implement Bluetooth link between our app and device, the role of Bluetooth client was assigned to the device while our app acts as the Bluetooth server. For communication between them, a baud rate of 9600 bps was set. Simple user operated buttons were placed on the app screen to easily connect and disconnect the link with the device. The app searches for all the available devices within a range and the oximeter device has to be selected manually. Once the device is connected and the required parameter is selected from the screen, the device calculates the value and transmission of data occurs from the oximeter to the app. VII. RESULTS AND COMPARISON Results were obtained and observed in different forms. At first, we directly tested the receiver circuit by obtaining the voltage values on an oscilloscope. The final output that we received was on our Android App. Fig 10. Results obtained on the Android App We compared the heart rate values of a person under various stages of physical activity as well as the performance of our device as compared to the ones used commercially. The heartbeat values measured by our device came close to the actual heartbeat values, as shown in Table 1 below. TABLE I. Actual Heart Rate HEART RATE COMPARISON Measured Heart Rate Device TABLE II. Commercially Used Oximeters PERFORMANCE COMPARISON SpO2 Accuracy +-1% No Phone App Cost This Device +-1.5% Yes $ 17 $59.32 (approx.) VIII. CONCLUSION The normal heart rate of healthy human being lies around 72 beats per minute with the complete range varying from 60 to 120 bpm. If the heart rate falls below 40bpm, it is an indication of bradycardia whereas if it rises above 150bpm, it indicates tachycardia [13]. The oxygen saturation level in blood of a healthy human being lies in the range of %. If this value slips below 94%, this indicates a deficiency of oxygen levels in the blood and a specialist should be contacted immediately. The device ran successfully after being interfaced with the designed Android App. The central idea of this device is to promote the use of such health monitoring devices among the population of the world which usually gets left out from being introduced to technological inventions in medical domain, primarily because of financial issues. Such low cost devices with considerable accuracy can help in bringing down the rate of several medical conditions such as Hypoxemia and Cardiovascular malformations, a defect which accounts for 6-10% of all infant deaths and 20-40% of deaths caused by congenital malformation [14] [15]. Furthermore, modules such as those used to calculate breathing rate [16], blood pressure, insulin level etc., can be designed with the same microcontroller so that this device can act as a one-stop health monitoring system for the user without inflating the cost of the system too much. Although at present the performance of the proposed system in this paper remains inferior to the ones available in the market at much higher costs, but with further research and analysis higher accuracy and precision for such low cost devices can be obtained in the future. REFERENCES [1] Martin, Lawrence (1999). All you really need to know to interpret arterial blood gases (2nd ed.). Philadelphia: Lippincott Williams & Wilkins. p. xxvi. ISBN [2] Eckman, Margaret (2010). Professional guide to pathophysiology (3rd ed.). Philadelphia: Wolters Kluwer/Lippincott Williams & Wilkins. p ISBN [3] Michiels, C. (2004). Physiological and Pathological Responses to Hypoxia. The American Journal of Pathology, 164(6), [4] Calvin K. Chan, Paul M. Vanhoutte, Hypoxia, vascular smooth muscles and endothelium, Acta Pharmaceutica Sinica B, Volume 3, Issue 1, February 2013, Pages 1-7, ISSN [5] Tavakoli, M., L. Turicchia, and R. Sarpeshkar. An Ultra-Low Power Pulse Oximeter Implemented with an Energy-Efficient Transimpedance Amplifier

6 Biomedical Circuits and Systems, IEEE Transactions on 4.1 (2010): Copyright 2010, IEEE [6] J. W. Severinghaus and J. F. Kelleher, Recent developments in pulse oximetry, Anesthesiology, vol. 76, no. 6, pp , 1992 [7] Hanning, C. D., and J. M. Alexander-Williams. Pulse Oximetry: A Practical Review. BMJ : British Medical Journal (1995): Print [8] A. Wongjan, A. Julsereewong, and P. Julsereewong, Continuous Measurements of ECG and SpO2 for Cardiology Information System, International MultiConference of Engineers and Computer Scientists 2009 Vol II IMECS 2009, March 18-20, 2009, Hong Kong [9] M. Nitzan and H. Taitelbaum, The measurement of oxygen saturation in arterial and venous blood," IEEE Instrumentation and Measurement Magazine, vol. 11, no. 3, pp. 9-15, June 2008 [10] R. M. Walls, M. F. Murphy, (2009), A comparison of transmittance and reflectance pulse oximetry during vascular surgery, Anesth Analg, Vol. 109, No. 6, PP, [11] S. Takatani, C.Davies, G.P. Noon, Y. Nose, M.E. DeBakey, A Noninvasive Reflectance Pulse Oximeter Sensor, Dept. of Surgery, Houston [12] N. Watthanawisuth, T. Lomas, A. Wisitsoraat, A.Tuantranont, Wireless Wearable Pulse Oximeter for Health Monitoring using Zigbee Wireless Sensor Network, National Electronics and Computer Technology Centre (NECTEC), Thailand [13] F. Adochiei, C. Rotariu, R. Ciobotariu and H. Costin, A Wireless Low- Power Pulse Oximetry System for Patient Telemonitoring, The 7 th International Symposium on Advanced Topics in Electrical Engineering, The faculty of Electrical Engineering, U.P.B., Bucharest, May 12-14, 2011 [14] de-wahl Granelli, Anne et al. Impact of Pulse Oximetry Screening on the Detection of Duct Dependent Congenital Heart Disease: A Swedish Prospective Screening Study in New borns. BMJ : British Medical Journal 338 (2009): a3037. PMC. [15] Ewer AK, Furmston AT, Middleton LJ, Deeks JJ, Daniels JP, Pattison HM, et.al. Pulse oximetry as a screening test for congenital heart defects in new born infants: a test accuracy study with evaluation of acceptability and costeffectiveness. Health Technol Assess 2012; 16(2) [16] W. S. Johnston, Y. Mendelson, Extracting Breathing Rate Information from a Wearable Reflectance Pulse Oximeter Sensor, Proceedings of the 26th Annual International Conference of the IEEE EMBS San Francisco, CA, USA [17] Das, Jyotirmoy, Amit Aggarwal, and Naresh Kumar Aggarwal. Pulse Oximeter Accuracy and Precision at Five Different Sensor Locations in Infants and Children with Cyanotic Heart Disease. Indian Journal of Anaesthesia 54.6 (2010): PMC. Web. 6 June