`The 5th International Summer School and Symposium on Medical Devices and Biosensors
`The Chinese University of Hong Kong, HKSAR, China. Jun 1-3, 2008
`
`A Wearable “Electronic Patch” for Wireless Continuous Monitoring of
`Chronically Diseased Patients
`
`Rasmus G. Haahr, Sune Duun, Erik V. Thomsen, Karsten Hoppe, and Jens Branebjerg
`
`Abstract— We present a wearable health system (WHS) for
`non-invasive and wireless monitoring of physiological signals.
`The system is made as an electronic patch where sensors,
`low power electronics, and radio communication are integrated
`in an adhesive material of hydrocolloid polymer making it a
`sticking patch. The patch is made with a reusable part and a
`disposable part which contains the adhesive material and the
`battery. This part is changed once every week. The patch has
`a size of 88 mm by 60 mm and a thickness of 5 mm. It is
`made for attachment on truncus or the greater muscle groups.
`The patch is demonstrated in two applications: Monitoring of
`electromyography (EMG) and arterial oxygen saturation by
`pulse oximetry (SpO2). The pulse oximetry sensor is made of a
`concentric backside Silicon photodiode with a hole in the middle
`for the two light sources. This makes it suitable for reflectance
`pulse oximetry. For the EMG application three standard dry
`silver electrodes are used separated by 10 mm.
`
`I. INTRODUCTION
`
`During the last decade there have been an increasing
`interest in new technology and innovative systems for the
`health care system. Significant factors such as limitations in
`the health care system’s resources, the aging population, and
`chronic conditions are motivating research.
`In this context non-invasive wearable health systems
`(WHS) for monitoring elderly and chronically diseased
`people outside hospitals have been developed. Various ap-
`proaches have taken place for continuously monitoring of
`vital signs. H. Asada et. al presented in 1998 a system
`made as a small finger ring sensor [1]. A European project
`“Sensation” have followed the same the approach [2]. J.
`Kang et. al have developed an instrument worn on the wrist
`[3]. Devices attached to truncus have not been made, but
`sensors have been integrated in textiles or clothes [4], [5].
`Following these ideas an European project “Wealthy” have
`developed an integrated system [6], [7].
`The vision for our research is to create a technology which
`allows for continuously monitoring of elderly and chronically
`diseased on a 24/7 basis. Furthermore, the sensor system
`must be able to function for a week without having to be
`changed. The technology should be convenient, versatile,
`easy to use, well integrated in the modern health care system
`and should provide safety and service for the patients.
`
`Manuscript received February 25, 2008. This work was party supported
`by the Danish Ministry of Science, Technology and Innovation.
`Rasmus G. Haahr, Sune Duun, and Erik V. Thomsen are with the Tech-
`nical University of Denmark, Department of Micro and Nanotechnology,
`Oersted Plads building 345 east, 2800 Kgs. Lyngby, Denmark (Phone:
`+ 45 45255700; Fax: +45 45887762; e-mail: rasmus.haahr@mic.dtu.dk,
`sdu@mic.dtu.dk, evt@mic.dtu.dk).
`Karsten Hoppe and Jens Branebjerg are with Delta A/S, Venlighedsvej
`4, 2970 Hørsholm, Denmark (e-mail: kh@delta.dk, jab@delta.dk).
`
`Fig. 1. The “Electronic Patch”. The patch has as a size of 88 mm by 60
`mm and is 5 mm thick.
`
`In this paper we report the recent development of a patient
`monitoring system the ”Electronic Patch” shown in Fig.
`1 which we previously have proposed [8], [9]. The patch
`can monitor various physiological signals, analyze these and
`when needed either transmit an alarm or continuously stream
`data over a wireless network.
`In this paper we describe the Electronic Patch in the
`application of monitoring electromyography (EMG), and
`pulse oximetry for monitoring the oxygen saturation (SpO2)
`and heart rate. For the latter case we have previously reported
`a design of a novel ring shaped photodiode [10] which have
`a concentric detection of light transmitted from two light
`emitting diodes (LEDs) placed in a hole in the center of the
`chip.
`
`II. POSSIBILITIES AND LIMITATIONS FOR WEARABLE
`AND NON-INVASIVE MONITORING
`
`Today clinical care and monitoring of elderly and chroni-
`cally diseased is typically done by frequent health examina-
`tions at the hospital or local physician. A new paradigm for
`patient monitoring where wearable and wireless systems are
`introduced to complete routine health examinations induce
`possibilities for monitoring health conveniently anywhere at
`all times. In non critical applications it can log physiological
`information and thereby improve treatment. In critical appli-
`cations it can automatically transmit an alarm and data if a
`medical incident is detected.
`
`A. Possibilities
`
`WHS are preferably non-invasive to limit infection risks.
`A variety of physiological data can be measured by non-
`invasive techniques e.g. electrocardiogram (ECG), elec-
`tromyogram (EMG) electroencephalogram (EEG), blood
`
`978-1-4244-2253-1/08/$25 ©2008 IEEE
`
`66
`
`Authorized licensed use limited to: Danmarks Tekniske Informationscenter. Downloaded on November 11, 2009 at 04:46 from IEEE Xplore. Restrictions apply.
`
`Apple Inc.
`APL1020
`U.S. Patent No. 8,989,830
`
`
`
`pressure, heart rate, photoplethysmogram (PPG), phonocar-
`diogram (PCG), oxygen saturation (SpO2), saturation of
`carbon monoxide (SpCO), skin temperature, respiration fre-
`quency, respiration volume, sweat production, tissue perfu-
`sion, and motion. There are several measuring principles that
`can be used for non-invasive monitoring of such physiolog-
`ical signals.
`1) Electrical: Electric potentials are used for ECG, EMG
`and EEG and used for monitoring the hearth and nervous
`systems. Impedance measurements can be used for monitor-
`ing the respiration frequency and resistance for monitoring
`the skin temperature.
`2) Optical: Optical measurements with non-coherent light
`is typically based on absorption for measuring PPGs,
`whereas coherent light can be used for laser Doppler blood-
`flow monitoring. PPGs have had an increased interest in the
`recent years due to potential of extracting more physiological
`parameters such as blood pressure [11], [12].
`3) Mechanical: Mechanical sensors can be used for mea-
`suring respiration volume, motion, blood pressure, phonocar-
`diogram (PCG), and ultrasound based measurements.
`4) Chemical: Chemical measurements are typically inva-
`sive and have therefore limited possibilities in non-invasive
`systems. However,
`there are electrochemical sensors that
`measure CO2 and O2 transcutaneously. Analysis on sweat
`is also possible, but this is not clinical practice today.
`
`B. Limitations
`
`In general WHS are limited by two main factors: Comfort
`and power consumption. An additional third factor is motion
`artefacts which is a well-know problem for some types of
`measurements e.g. ECG and PPG.
`For WHS to work on a 24/7 basis attached to people
`living an everyday life the attachment must neither be
`physical nor visual annoying for the person so that the person
`is comfortable wearing it. WHS should therefore be bio-
`compatible and must not cause pressure to the tissue which
`will prevent blood circulation.
`The size of the systems is obviously also a limiting
`factor and in general WHS should be as small as possible
`and follow natural body shapes to provide most comfort.
`In the case WHS are attached directly to the body they
`must therefore be as flat as possible and not thicker than a
`few millimeters. Alternatively, WHS could be integrated in
`things that people would wear anyway e.g. clothes. Typical
`locations for attaching WHS are the finger or wrist e.g. [1],
`[3]; however, these locations are not convenient in daily
`activities. Locations on truncus hidden by clothes seems
`more attractive. These locations have previously been utilized
`in systems where sensors are incorporated in clothes [4], but
`this approach does not offer a firm attachment of the sensors
`directly to the body.
`The power consumption for WHS is a critical issue and
`is limited either by the battery or by the amount of energy
`that can be harvested from the body e.g. thermoelectrically.
`Since the size of WHS should be as small as possible the
`systems cannot have large batteries and due to convenience
`
`Ring shaped photodiode with LEDs in the center mounted on
`Fig. 2.
`bottom side of PCB.
`
`they should last as long as possible. Power is consumed by
`the sensors, frontend electronics, digital processing of data,
`and radio communication.
`Sensors for some physiological signals require very little
`power e.g. potential measurements as EMG. But other physi-
`ological signals require more energy consuming sensors e.g.
`pulse oximetry due the driving current for the LEDs. For
`some applications it is therefore necessary to focus both on
`the energy consumed by the sensors as well as the successive
`frontend electronics and digital signal processing.
`Power management can be done by several approaches.
`In general radio communication should be limited since this
`is highly power consuming. It is therefore advantageous to
`have signal analysis done in the WHS and then only transmit
`data or just an alarm when abnormalities arise.
`The nature of the physiological signals should also be
`considered. Some change in the scale of hours e.g. tem-
`perature some change in the scale of minutes e.g. oxygen
`saturation and some change in the scale of seconds e.g. the
`status of the heart as monitored by ECG. The requirement to
`the measuring frequency is therefore different between these
`types of signals. By measuring quasi continuously with a
`period corresponding to the natural period of the change in
`the signals power can be conserved without losing critical
`information.
`
`III. OVERVIEW OF THE ELECTRONIC PATCH SYSTEM
`
`Based on the following considerations we have developed
`a WHS as an electronic patch. The Electronic Patch is
`a genuine platform which is compatible with many types
`of sensors. In this paper we describe two applications:
`monitoring of EMG and SpO2 by pulse oximetry. The EMG
`sensor is intended for detection of convulsions during sleep
`and the pulse oximetry sensor is intended for people suffering
`from heart disorders, chronical lung diseases (COLD), sleep
`apnea, and professionals during work such as fire fighters.
`The Electronic Patch consists of a printed circuited board
`(PCB) where sensors are mounted on the bottom, and the
`top contains all the electronics and radio communication.
`The PCB is encapsulated in a hard plastic box and attached
`to the body by an adhesive material of hydrocolloid polymer.
`
`A. Sensors
`
`The EMG sensor have a standard design made by three
`silver electrodes distributed evenly on the PCB with a
`separation of 10 mm. The pulse oximetry sensor comprises
`a concentric photodiode with two LEDs in the middle a red
`
`Authorized licensed use limited to: Danmarks Tekniske Informationscenter. Downloaded on November 11, 2009 at 04:46 from IEEE Xplore. Restrictions apply.
`
`67
`
`
`
`Crystal, 32 MHz
`
`LEDs current control, I2C protocol
`
`MCU and RF
`
`Analog-digital converter
`
`Op.amps. (photodiode)
`
`Memory
`
`Op.amp. (thermistor)
`
`Fig. 3. Top side of the PCB showing the types of electronic components
`which is utilized in the pulse oximetry version of the Electronic Patch.
`
`(660 nm) and infrared (940 nm) [10]. The sensor is shown
`in Fig. 2.
`
`B. Electronics
`
`The top side of the PCB contains the electronics as shown
`in Fig. 3. It contains analog frontend electronics, a low
`power microprocessor with a built-in radio, and memory. The
`microprocessor uses from 190 µA at 32 kHz with the radio
`off to 27 mA at 32 MHz with the radio on. The power usage
`of the microprocessor will be application dependent. In the
`pulse oximetry sensor we also have an I2C current controller
`to control the LEDs. The patch is powered by a coin size 3
`V Lithium-ion battery with 170 mAh.
`
`C. Wireless communication and network
`
`The wireless networking in the Electronic Patch is based
`on a 2.4 GHz radio and a proprietary protocol which allows
`the patch to work in a wireless personal area network,
`but not as an independent system in direct contact with
`service providers or hospitals. However, this contact can be
`made by external access points connected to the internet
`e.g. smart phones. Access points could also be installed
`in the person’s home or other daily environments. The
`advantage using this solution is that power consuming long
`distance communication is placed outside the patch. This
`configuration also supports the service of many patches.
`For instance in the case of assisted living homes where
`many elderly could be monitored by individual patches each
`connected to the same network of access points covering
`the entire estate. A proprietary protocol has been employed
`instead of the ZigBee and Bluetooth protocols due to lower
`power consumption. The drawback is a limited range of a
`few meters. This would be increased by using the Bluetooth
`protocol.
`
`D. Mechanical assembly
`
`The mechanical assembly is shown in Fig. 4 and the final
`patch with the pulse oximetry sensor is shown in Fig. 5.
`Sensors and electronics are encapsulated in a bio-compatible
`plastic housing which protect the electronics from sweat and
`moisture. The pulse oximetry sensor is further protected by
`a transparent membrane and the EMG sensor has an epoxy
`seal. With this solution the system can even be warn during
`a shower.
`The patch comes in two parts: 1) A reusable sensor part
`consisting of a bottom- (f) and middle plastic housing (d),
`
`(a) Adhesive patch
`
`(b) Plastic housing - top
`
`(c) Battery
`
`(d) Plastic housing- middle
`
`(e) Printed circuit board, PCB
`
`(f) Plastic housing - bottom
`
`(g) Bio-compatible “window”
`
`Fig. 4. CAD drawing of the parts in the electronic patch and how they
`are assembled.
`
`The assembled patch with a pulse oximetry sensor made as a
`Fig. 5.
`concentric photodiode around two LEDs placed in the center. The little
`square frame around the LEDs is to prevent light going directly from the
`LEDs into the photodiode.
`
`sensors and electronics (e). 2) A disposable part consisting
`of the adhesive patch (a), top housing (b), and battery (c).
`The adhesive patch has to be changed once every week due
`to dead skin cells. This is therefore the period which the
`battery has been designed to last. The adhesive patch is
`designed for attaching the plastic housing onto the skin and
`the hydrocolloid polymer allows for diffusion of moisture
`away from the skin.
`
`IV. EMG APPLICATION
`
`Electromyography is a method of detecting muscle activ-
`ity. The methods relies on the change of membrane potential
`of the muscle cells with muscle activity. The resting muscle
`cell have a potential across the cell membrane of approxi-
`mately -90 mV. During muscle activity the membrane poten-
`tial change to approximately 15 mV. This can occur both in
`spikes when the muscle is stimulated or constantly when
`the muscle contraction is tetanic. EMG can be measured
`both non-invasively on the skin surface above the muscle
`or invasively by needles.
`We have used a standard configuration for surface EMG
`where the potential
`is measured between two electrodes
`relative to a third electrode placed in between. The measured
`signal is amplified, and to save power an analog circuit for
`detection of spikes have been employed. The microprocessor
`is then only turned on whenever spikes are detected and
`the muscle is active. This is demonstrated in Fig. 6 where
`the yellow (top) curve is the recorded EMG signal and
`
`Authorized licensed use limited to: Danmarks Tekniske Informationscenter. Downloaded on November 11, 2009 at 04:46 from IEEE Xplore. Restrictions apply.
`
`68
`
`
`
`Current[µA],IRLEDon
`
`63.4
`
`63.2
`
`63.0
`
`62.8
`
`Red LED on
`IR LED on
`
`34.3
`
`34.2
`
`34.1
`
`34.0
`
`Current[µA],redLEDon
`
`0
`
`1
`
`2
`
`3
`Time (s)
`
`4
`
`5
`
`Photoplethysmograms measured with Electronic Patch pulse
`Fig. 7.
`oximetry sensor
`
`has the best signal to noise ratio. One such ring sensor is
`seen in Fig. 5.
`To ease the assembly we have chosen to make backside
`photodiodes which have the junction and both contacts on the
`side facing the PCB. Therefore, no wirebonding is necessary.
`To shield from ambient light and to optimize transmission
`at the two wavelengths of interest i.e. 660 nm and 940 nm a
`two layer antireflection filter consisting of 550 nm PECVD
`Silicon Nitride on 50 nm thermal dry Silicon Oxide has
`been employed. This filter reach optical transmission > 98%
`at 660 nm and 940 nm and suppressing other wavelengths
`to approximately 50 % in the range 600 nm - 1100 nm.
`For wavelengths below 600 nm the tissue absorbtion is very
`strong and hence ambient light at these wavelengths does not
`course problems. The photodiodes are also patterned with
`Aluminum on the side of the light entrance to give a well-
`defined area of light gathering.
`Photopletsysmograms like the one shown in Fig. 7 have
`been recorded with the fabricated sensor. From the PPGs the
`pulse and the oxygen saturation can be calculated.
`To further optimize the power consumption of the pulse
`oximetry sensor the duty cycle of the LEDs, DLED, can be
`considered. The minimal duty cycle that is possible, when at
`least 95 % of the LED power must be maintained, is given by
`the sampling frequency and the bandwidth of the photodiode
`amplifier circuit [15]. In our case we get
`
`DLED ≃ 2 · fs/BW = 2 · 1kHz/4kHz = 50%
`
`(2)
`
`When lit the LEDs typically use 20 mA at 1.5 V. The I2C
`current controller needs 10 mA at 3 V to deliver 20 mA at
`1.5 V. Having a duty cycle of 50% on the LEDs the I2C
`current controller on average will use 5 mA at 3 V. If we
`measure continuously the LEDs alone would use the battery
`in 34 hours. Therefore, we would like to reduce the LED
`power consumption by at least a factor of 10. Because then
`we can measure continuously for a week and only use 85
`mAh or half the battery power available on the LEDs. One
`way to do this will be to improve the speed of the photodiode
`amplifier circuit by lowering the photodiode capacitance.
`
`VI. DISCUSSION
`
`The Electronic Patch have at this time been tested in the
`laboratory and on persons wearing it for periods of one week.
`
`The yellow (top) graph shows a recorded EMG signal by the
`Fig. 6.
`Electronic Patch. The blue (bottom) curve indicates the turn on and off of
`the microprocessor which is controlled by an analog frontend spike detecting
`circuit.
`
`the blue (lower) curve indicates the turn on and off of the
`microprocessor. The microprocessor then analyze the EMG
`signal and evaluate if convulsions are taking place.
`
`V. PULSE OXIMETRY APPLICATION
`
`A pulse oximetry sensor detects pulse and arterial oxygen
`saturation. It is an optical technique invented by T. Aoyagi in
`1972 [13] and is based on absorption changes of light with
`the blood flow. Pulse oximetry relies on the difference in the
`absorption spectra between oxygenated haemoglobin (HbO2)
`and deoxygenated haemoglobin (Hb). In [9] it is shown that
`the ratio between absorption coefficients of HbO2 and Hb
`makes wavelengths of 660 nm and 940 nm suitable.
`For the pulse oximetry application we have chosen to cus-
`tom design pn Silicon photodiodes. This allows for optimiza-
`tion of the photodiodes for the pulse oximetry application.
`To minimize the necessary driving current of the LEDs we
`have fabricated large area photodiodes which are concentric
`around the LEDs and hence optimized for collection of
`backscattered light from the tissue [10]. This approached was
`pioneered by Y. Mendelson [14]; however, with the use of
`several discrete photodiodes. The photodiodes have a chip
`size of 14 mm by 14 mm and with various active areas
`ranging from 22 mm2 to 121 mm2. This area is up to 20
`times larger than what is used in a Nellcor wired reflectance
`sensor. The largest photodiode is shown in Fig. 2. We have
`described the fabrication process elsewhere [10].
`Increasing the photodiode area also increases the capaci-
`tance and this will lower the speed of the photodiode, hence
`there is a tradeoff between photodiode area and speed. In
`our system we use a sampling rate, fs, of 1 kHz. The
`capacitances of the largest photodiodes are 24 nF ±2 nF.
`Given a photodiode transimpedance amplifier circuit with a
`104 amplification the bandwidth, BW , will approximately
`be given by:
`
`BW ≃ (CPD · RAmp)−1 = (24nF · 10kΩ)−1 = 4 kHz (1)
`
`We have fabricated several 1 mm wide rings with radii
`from 3.5 mm to 6.5 mm. This is done to gain knowledge
`about at what radii on a specific body location the signal
`
`Authorized licensed use limited to: Danmarks Tekniske Informationscenter. Downloaded on November 11, 2009 at 04:46 from IEEE Xplore. Restrictions apply.
`
`69
`
`
`
`Tests in a clinical setting have not been done at this time.
`Thus there are still questions to be examined regarding long-
`term use and actual benefits compared to traditional health
`monitoring and other wearable health monitoring systems.
`Likewise clinical experiments is also required to calibrate
`the pulse oximeter. However, this is a standard procedure.
`In contrast to other systems attached around the finger,
`wrist or integrated in clothes the Electronic Patch is mounted
`directly to the truncus or greater muscles groups offering a
`very firm attachment of the sensors and no wiring between
`several units. The Electronic Patch is therefore a versatile
`system and very simple to use.
`The reliability of WHS are both dependent on the attach-
`ment of the sensors and the algorithms used to analyze the
`signals. Development of algorithms which provide enough
`and reliable artificial intelligence are therefore important.
`The most suitable locations to measure depend on the
`application. For the EMG application the location must be
`the on muscle which is being examined. Typically, this will
`be the larger muscle groups e.g. biceps. In case of the pulse
`oximetry application further research have to be done in this
`direction. The Electronic Patch is not compatible with the
`typical locations for pulse oximetry i.e. the finger or forehead
`since this will be visually annoying for the patient. We are
`targeting a location on truncus either ventral or dorsal and
`we are currently investigating this. Pulse oximetry on the
`sternum have previously been demonstrated by [16], but it
`needs more investigation before it can be used clinically.
`The most suitable wireless network to use in WHS de-
`pendent on the application. If WHS are intended for contin-
`uously monitoring the WHS should be able to get online
`often and the power consumed by the radio is therefore
`important. However, in the applications where WHS only
`should transmit an alarm when a critical condition arises it
`is more important that the WHS can get online. Since it
`only need to transmit once it can use more power and in this
`case the global system for mobile communications (GSM)
`network could be employed.
`As we mention in this paper the large capacitances of
`the large area photodiodes impose a problem regarding the
`speed. This could be improved by pin photodiodes. Our
`fabrication process for the photodiodes are compatible with
`a step where an epitaxial layer of Silicon is grown making
`the photodiodes of the pin type.
`
`VII. CONCLUSION
`
`We have developed an Electronic Patch as a wearable
`health monitoring system. The Electronic Patch is a genuine
`platform which can be used with many types of sensors.
`The Electronic Patch is demonstrated in two very different
`applications EMG and pulse oximetry monitoring. These
`applications rely on two different measuring principles i.e.
`electrical and optical.
`For the pulse oximetry application a novel concentric
`photodiode is used which lower the requirement to the LED
`driving current. For the EMG application standard silver
`electrodes are used. The electronics in the Electronic Patch
`
`are low power surface mountable components. The patch is
`powered by a 3 V Lithium-ion battery which last a week and
`is changed when the adhesive material needs to be changed
`anyway.
`We will fabricate pin photodiodes in order to reduce
`the capacitance of the photodiodes which will allow for a
`lower duty cycle of the LEDs. This will reduce the power
`consumption of the pulse oximetry sensor.
`The Electronic Patch is made compatible with wireless
`personal area network; however, the patch also relies on
`external devices for online communication with service
`providers or hospitals. The EMG and pulse oximetry ap-
`plication sensors were shown to work, but clinical test and
`trials of the Electronic Patch still remain to be investigated.
`
`REFERENCES
`
`[1] H. Asada, “The Ring Sensor: a New Ambulatory Wearable Sensor
`for Twenty-Four Hour Patient Monitoring”, Proc. 20th Annual IEEE
`EMBS conf., vol. 4, pp 1906-1909, 1998
`[2] J. Sola el. al, “SpO2 Sensor Embedded in a Finger Ring: desing and
`implementation” Proc. 28th Annual IEEE EBMS conf. , pp 4295-4298,
`2006
`[3] J. M. Kang, T. Yoo, H. C. Kim, “A wrist-worn integrated health mon-
`itoring instrument with a tele-reporting device for telemedicine and
`telecare”, IEEE Transactions on Instrumentation and Measurement,
`vol. 55, nr. 5, pp 1655-1661, 2006
`[4] A. D. Santa, A. Mazzoldi, D. De Rossi, “Dressware: warable hard-
`ware” Mater Sci Eng, C7:31.37, 1999
`[5] G. Loriga, N. Taccini, D. De Rossi, R. Paradiso, “Textile Sensing
`Interfaces for Cardiopulmonary Signs Monitoring” Proc. 27th Annual
`IEEE EBMS conf. , pp 7349-7352, 2005
`[6] R. Paradiso, G. Loriga, N. Taccini, “Wearable System for Vital
`Signs Monitoring” Wearable eHealth Systems for Personalised Health
`Management, pp 253-259, 2004
`[7] R. Paradiso, G. Loriga, N. Taccini, “A Wearable Health Care System
`Based on Knitted Integrated Sensors” IEEE Transactions on Informa-
`tion Technology in Biomedicine, vol. 9, no. 3, pp 337-344, 2005
`[8] E. V. Thomsen, A. Hyldg˚ard, K. Birkelund, R. G. Haahr, S. Duun, R.
`Elsubaihi, J. Branebjerg, “Distributed sensors: Applciations, Fabrica-
`tion and Challenges” Proc. 20th Eurosensors Conf., 2006.
`[9] R. G. Haahr, S. Duun, K. Birkelund, P. Raahauge, P. Petersen, H. Dam,
`L. Nørgaard, E. V. Thomsen, “A Novel Photodiode for Reflectance
`Pulse Oximetry in low-power applications” Proc. 29th Annual IEEE
`EMBS conf., pp. 2350 - 2353, 2007
`[10] S. Duun, R. G. Haahr, K. Birkelund, P. Raahauge, P. Petersen, H. Dam,
`L. Nørgaard, E. V. Thomsen, “A Novel Ring Shaped Photodiode for
`refelctance Pulse Oximetry in Wireless Applications” Proc. 6th Annual
`IEEE Sensors Conf., pp. 596 - 599, 2007
`[11] X. F. Teng, Y. T. Zhang, “Continous and Noninvasive Estimation of
`Arterial Pressure Using a Photoplethysmographic Approach” Proc.
`25th Annual IEEE EBMS conf. , pp 3153-3156, 2003
`[12] P. A. Shaltis, A. Reisner, H. H. Asada, “Wearable, Cuff-less PPG-
`Based Blood Pressure Monitor with Novel Height Sensor” Proc. 28th
`Annual IEEE EBMS conf. , pp 908-911, 2006
`[13] T. Aoyagi, M. Kishi, K. Yamaguchi, S. Watanabe. “Improvement of
`the earpiece oximeter”, Abstracts of the 13th Annual Meeting of the
`Japanese Society of Medical Electronics and Biological Engineering,
`pp 90-91, 1974 (Jap)
`[14] Y. Mendelson, “Design and evaluation of a new reflectance pulse
`oximeter sensor”, Medical Instrumentation , vol. 22, nr. 4, pp 167-
`173, 1988
`[15] E. A. Pel´aez and E. R. Villegas, “LED power reduction trade-offs for
`ambulatory pulse oximetry”Proc. 29th Annual IEEE EMBS conf., pp.
`2296 - 2299, 2007.
`[16] J. Sola, O. Chetelat, J. Krauss, “One the reliability of pulse oximery
`at the sternum” Proc. 29th Annual IEEE EBMS conf. , pp 1537, 2007
`
`Authorized licensed use limited to: Danmarks Tekniske Informationscenter. Downloaded on November 11, 2009 at 04:46 from IEEE Xplore. Restrictions apply.
`
`70