`EMBS Annual International Conference
`New York City, USA, Aug 30-Sept 3, 2006
`
`ThB14.4
`
`A Wearable Reflectance Pulse Oximeter for Remote Physiological
`Monitoring
`Y. Mendelson*, Member, IEEE, R. J. Duckworth, Member, IEEE, and G. Comtois, Student Member, IEEE
`
`Abstract—To save life, casualty care requires that trauma
`injuries are accurately and expeditiously assessed in the field.
`This paper describes the initial bench testing of a wireless
`wearable pulse oximeter developed based on a small forehead
`mounted sensor. The battery operated device employs a
`lightweight optical reflectance sensor and incorporates an
`annular photodetector to reduce power consumption. The
`system also has
`short range wireless communication
`capabilities to transfer arterial oxygen saturation (SpO2), heart
`rate (HR), body acceleration, and posture information to a
`PDA. It has the potential for use in combat casualty care, such
`as for remote triage, and by first responders, such as
`firefighters.
`
`S
`
`INTRODUCTION
`I.
`TEADY advances in noninvasive physiological sensing,
`hardware miniaturization, and wireless communication
`leading
`to
`the development of new wearable
`are
`technologies that have broad and important implications for
`civilian and military applications [1]-[2]. For example, the
`emerging development of compact, low-power, small-size,
`light- weight, and unobtrusive wearable devices may
`facilitate remote noninvasive monitoring of vital signs from
`soldiers during training exercises and combat. Telemetry of
`physiological information via a short-range wirelessly-linked
`personal area network can also be useful for firefighters,
`hazardous material workers, mountain climbers, or
`emergency first-responders operating in harsh and hazardous
`environments. The primary goals of such a wireless mobile
`platform would be to keep track of an injured person’s vital
`signs, thus readily allowing the telemetry of physiological
`information to medical providers, and support emergency
`responders in making critical and often life saving decisions
`in order to expedite rescue operations. Having wearable
`physiological monitoring could offer far-forward medics
`numerous advantages, including the ability to determine a
`casualty’s condition remotely without exposing the first
`
`Manuscript received April 3, 2006. This work is supported by the U.S.
`Army Medical Research and Material Command under Contract No.
`DAMD17-03-2-0006. The views, opinions and/or findings are those of the
`author and should not be construed as an official Department of the Army
`position, policy, or decision, unless so designated by other documentation.
`*Corresponding author – Y. Mendelson is a Professor in the Department
`of Biomedical Engineering, Worcester Polytechnic Institute, Worcester,
`MA 01609 USA (phone: 508-831-5103; fax: 508-831-5541; e-mail:
`ym@wpi.edu).
`R. J. Duckworth is a Professor in the Department of Electrical and
`Computer Engineering, Worcester Polytechnic Institute, Worcester, MA
`01609 USA (rjduck@ece.wpi.edu).
`G. Comtois is a M. S. student in the Department of Biomedical
`Engineering, Worcester Polytechnic Institute, Worcester, MA 01609 USA
`(comtoisg@wpi.edu).
`
`the
`identifying
`increased risks, quickly
`to
`responders
`severity of injuries especially when the injured are greatly
`dispersed over large geographical terrains and often out-of-
`site, and continuously tracking the injured condition until
`they arrive safely at a medical care facility.
`Several technical challenges must be overcome to address
`the unmet demand for long-term continuous physiological
`monitoring in the field. In order to design more compact
`sensors and improved wearable instrumentation, perhaps the
`most critical challenges are to develop more power efficient
`and
`low-weight devices. To become effective,
`these
`technologies must also be robust, comfortable to wear, and
`cost-effective. Additionally, before wearable devices can be
`used effectively in the field, they must become unobtrusive
`and should not hinder a person’s mobility. Employing
`commercial off-the-shelf (COTS) solutions, for example
`finger pulse oximeters to monitor blood oxygenation and
`heart rate, or standard adhesive-type disposable electrodes
`for ECG monitoring, is not practical for many field
`applications because they limit mobility and can interfere
`with normal tasks.
`A potentially attractive approach to aid emergency
`medical teams in remote triage operations is the use of a
`wearable pulse oximeter to wirelessly transmit heart rate
`(HR) and arterial oxygen saturation (SpO2) to a remote
`location. Pulse oximetry is a widely accepted method that is
`used for noninvasive monitoring of SpO2 and HR. The
`method is based on spectrophotometric measurements of
`changes in the optical absorption of deoxyhemoglobin (Hb)
`Noninvasive
`(HbO2).
`and
`oxyhemoglobin
`spectrophotometric measurements of SpO2 are performed in
`the visible (600-700nm) and near-infrared (700-1000nm)
`spectral regions. Pulse oximetry also relies on the detection
`of photoplethysmographic (PPG) signals produced by
`variations in the quantity of arterial blood that is associated
`with periodic contractions and relaxations of the heart.
`Measurements can be performed in either transmission or
`reflection modes. In transmission pulse oximetry, the sensor
`can be attached across a fingertip, foot, or earlobe. In this
`configuration,
`the
`light emitting diodes (LEDs) and
`photodetector (PD) in the sensor are placed on opposite sides
`of a peripheral pulsating vascular bed. Alternatively, in
`reflection pulse oximetry, the LEDs and PD are both
`mounted side-by-side on the same planar substrate to enable
`readings from multiple body
`locations where
`trans-
`illumination measurements are not feasible. Clinically,
`forehead reflection pulse oximetry has been used as an
`alternative approach to conventional transmission-based
`
`1-4244-0033-3/06/$20.00 ©2006 IEEE.
`
`912
`
`1
`
`APPLE 1010
`
`
`
`system can be programmed to alert on alarm conditions,
`such as sudden trauma, or physiological values out of their
`normal range. It also has the potential for use in combat
`casualty care, such as for remote triage, and for use by first
`responders, such as firefighters.
`Key features of this system are small-size, robustness, and
`low-power consumption, which are essential attributes of
`wearable physiological devices, especially for military
`applications. The system block diagram (Fig. 2), is described
`in more detail below.
`
`oximetry when peripheral circulation to the extremities is
`compromised.
`Pulse oximetry was initially intended for in-hospital use
`on patients undergoing or recovering from surgery. During
`the past few years, several companies have developed
`smaller pulse oximeters, some including data transmission
`via telemetry, to further expand the applications of pulse
`oximetry. For example, battery-operated pulse oximeters are
`now attached to patients during emergency transport as they
`are being moved from a remote location to a hospital, or
`between hospital wards. Some companies are also offering
`smaller units with improved electronic filtering of noisy
`PPG signals.
`Several reports described the development of a wireless
`pulse oximeter that may be suitable for remote physiological
`monitoring
`[3]-[4]. Despite
`the steady progress
`in
`miniaturization of pulse oximeters over the years, to date,
`the most significant limitation is battery longevity and lack
`of telemetric communication. In this paper, we describe a
`prototype forehead-based reflectance pulse oximeter suitable
`for remote triage applications.
`
`II. SYSTEM ARCHITECTURE
`The prototype system, depicted in Fig. 1, consists of a
`body-worn pulse oximeter that receives and processes the
`PPG signals measured by a small (φ = 22mm) and
`lightweight (4.5g) optical reflectance transducer. The system
`
`Fig. 2. System block diagram of the wearable, wireless, pulse oximeter.
`Sensor Module (top), Receiver Module (bottom).
`
`Sensor Module: The Sensor Module contains analog signal
`processing circuitry, ADC, an embedded microcontroller,
`and a RF transceiver. The unit is small enough so the entire
`module can be integrated into a headband or a helmet. The
`unit is powered by a CR2032 type coin cell battery with
`220mAh capacity, providing at least 5 days of operation.
`Receiver Module: The Receiver Module contains an
`embedded microcontroller,
`RF
`transceiver
`for
`communicating with the Sensor Module, and a Universal
`Asynchronous Receive Transmit (UART) for connection to
`a PC. Signals acquired by the Sensor Module are received by
`the embedded microcontroller which synchronously converts
`the corresponding PD output to R and IR PPG signals.
`Dedicated software is used to filter the signals and compute
`SpO2 and HR based on the relative amplitude and frequency
`content of the reflected PPG signals. A tri-axis MEMS
`accelerometer detects changes in body activity, and the
`information obtained through the tilt sensing property of the
`accelerometer is used to determine the orientation of the
`person wearing the device.
`To
`facilitate bi-directional wireless communications
`between the Receiver Module and a PDA, we used the
`DPAC Airborne™ LAN node module (DPAC Technologies,
`Garden Grove, CA). The DPAC module operates at a
`frequency of 2.4GHz, is 802.11b wireless compliant, and has
`a relatively small (1.6 × 1.17 × 0.46 inches) footprint. The
`wireless module runs off a 3.7VDC and includes a built-in
`
`Fig. 1. (Top) Attachment of Sensor Module to the skin; (Bottom)
`photograph of the Receiver Module (left) and Sensor Module (right).
`consists of three units: A Sensor Module, consisting of the
`optical transducer, a stack of round PCBs, and a coin-cell
`battery. The information acquired by the Sensor Module is
`transmitted wirelessly via an RF link over a short range to a
`body-worn Receiver Module. The data processed by the
`Receiver Module can be transmitted wirelessly to a PDA.
`The PDA can monitor multiple wearable pulse oximeters
`simultaneously and allows medics
`to collect vital
`physiological information to enhance their ability to extend
`more effective care to those with the most urgent needs. The
`
`913
`
`2
`
`
`
`TCP/IP stack, a radio, a base-band processor, an application
`processor, and software for a “drop-in” WiFi application. It
`has the advantage of being a plug-and-play device that does
`not require any programming and can connect with other
`devices through a standard UART.
`PDA: The PDA was selected based on size, weight, and
`power consumption. Furthermore, the ability to carry the
`user interface with the medic also allows for greater
`flexibility during deployment. We chose the HP iPAQ h4150
`PDA because it can support both 802.11b and Bluetooth™
`wireless communication. It contains a modest amount of
`storage and has sufficient computational resources for the
`intended application. The use of a PDA as a local terminal
`also provides a low-cost touch screen interface. The user-
`friendly
`touch screen of
`the PDA offers additional
`flexibility. It enables multiple controls to occupy the same
`physical space and the controls appear only when needed.
`Additionally, a touch screen reduces development cost and
`time, because no external hardware is required. The data
`from the wireless-enabled PDA can also be downloaded or
`streamed to a remote base station via Bluetooth or other
`wireless communication protocols. The PDA can also serve
`to temporarily store vital medical information received from
`the wearable unit.
`A dedicated National Instruments LabVIEW program was
`developed to control all interactions between the PDA and
`the wearable unit via a graphical user interface (GUI). One
`part of the LabVIEW software is used to control the flow of
`information through the 802.11b radio system on the PDA.
`A number of LabVIEW VIs programs are used to establish a
`connection, exchange data, and close
`the connection
`between the wearable pulse oximeter and the PDA. The
`LabVIEW program interacts with the Windows CE™
`drivers of the PDA’s wireless system. The PDA has special
`drivers provided by the manufacturer that are used by
`Windows CE™
`to
`interface with
`the 802.11b radio
`hardware. The LabVIEW program interacts with Windows
`CE™ on a higher level and allows Windows CE™ to handle
`the drivers and the direct control of the radio hardware.
`The user interacts with the wearable system using a
`simple GUI, as depicted in Fig. 3.
`
`Fig. 3. Sample PDA Graphical User Interface (GUI).
`The GUI was configured to present the input and output
`information to the user and allows easy activation of various
`
`functions. In cases of multiple wearable devices, it also
`allows the user to select which individual to monitor prior to
`initiating the wireless connection. Once a specific wearable
`unit is selected, the user connects to the remote device via
`the System Control panel that manages the connection and
`sensor control buttons. The GUI also displays the subject’s
`vital signs, activity level, body orientation, and a scrollable
`PPG waveform that is transmitted by the wearable device.
`The stream of data received from the wearable unit is
`distributed to various locations on the PDA’s graphical
`display. The most prominent portion of the GUI display is
`the scrolling PPG waveform, shown in Fig. 3. Numerical
`SpO2 and HR values are displayed is separate indicator
`windows. A separate tri-color indicator is used to annotate
`the subject’s activity level measured by the wearable
`accelerometer. This activity level was color coded using
`green, yellow, or red to indicate low or no activity, moderate
`activity, or high activity, respectively. In addition, the
`subject’s orientation is represented by a blue indicator that
`changes orientation according to body posture. Alarm limits
`could be set to give off a warning sign if the physiological
`information exceeds preset safety limits.
`One of the unique features of this PDA-based wireless
`system architecture is the flexibility to operate in a free
`roaming mode. In this ad-hoc configuration, the system’s
`integrity depends only on the distance between each node.
`This allows the PDA to communicate with a remote unit that
`is beyond the PDA’s wireless range. The ad-hoc network
`would
`therefore allow medical personnel
`to quickly
`distribute sensors
`to multiple causalities and begin
`immediate triage, thereby substantially simplifying and
`reducing deployment time.
`Power Management: Several features were incorporated
`into the design in order to minimize the power consumption
`of the wearable system. The most stringent consideration
`was the total operating power required by the Sensor
`Module, which has to drive the R and IR LEDs, process the
`data, and transmit this information wirelessly to the Receive
`Module. To keep the overall size of the Sensor Module as
`small as, it was designed to run on a watch style coin-cell
`battery.
` It should be noted that low power management without
`compromising signal quality is an essential requirement in
`optimizing
`the design of wearable pulse oximeter.
`Commercially available transducers used with transmission
`and reflection pulse oximeters employ high brightness LEDs
`and a small PD element, typically with an active area
`ranging between 12 to 15mm2. One approach to lowering the
`power consumption of a wireless pulse oximeter, which is
`dominated by the current required to drive the LEDs, is to
`reduce the LED duty cycle. Alternatively, minimizing the
`drive currents supplied to the R and IR LEDs can also
`achieve a significant reduction in power consumption.
`However, with reduced current drive, there can be a direct
`impact on the quality of the detected PPGs. Furthermore,
`since most of the light emitted from the LEDs is diffused by
`the skin and subcutaneous tissues, in a predominantly
`forward-scattering direction, only a small fraction of the
`incident light is normally backscattered from the skin. In
`
`914
`
`3
`
`
`
`addition, the backscattered light intensity is distributed over
`a region that is concentric with respect to the LEDs.
`Consequently, the performance of reflectance pulse oximetry
`using a small PD area is significantly degraded. To
`overcome this limitation, we showed that a concentric array
`of either discrete PDs, or an annularly-shaped PD ring, could
`be used to increase the amount of backscattered light
`detected by a reflectance type pulse oximeter sensor [5]-[7].
`Besides a low-power consuming sensor, afforded by
`lowering the driving currents of the LEDs, a low duty cycle
`was employed to achieve a balance between low power
`consumption and adequate performance. In the event that
`continuous monitoring is not required, more power can be
`conserved by placing the device in an ultra low-power
`standby mode. In this mode, the radio is normally turned off
`and is only enabled for a periodic beacon to maintain
`network association. Moreover, a decision to activate the
`wearable pulse oximeter can be made automatically in the
`event of a patient alarm, or based on the activity level and
`posture
`information
`derived
`from
`the
`on-board
`accelerometer. The wireless pulse oximeter can also be
`activated or deactivated remotely by a medic as needed,
`thereby further minimizing power consumption.
`
`III. IN VIVO EVALUATIONS
`Initial laboratory evaluations of the wearable pulse
`oximeter
`included
`simultaneous HR
`and
`SpO2
`measurements. The Sensor Module was positioned on the
`forehead using an elastic headband. Baseline recordings
`were made while the subject was resting comfortably and
`breathing at a normal tidal rate. Two intermittent recordings
`were also acquired while the subject held his breath for
`about 30 seconds. Fig. 4 displays about 4 minutes of SpO2
`and HR recordings acquired simultaneously by the sensor.
`
`Fig. 4. Typical HR (solid line) and SpO2 (dashed line) recording of two
`voluntary hypoxic episodes.
`The pronounced drops in SpO2 and corresponding increases
`in HR values coincide with the hypoxic events associated
`with the two breath holding episodes.
`
`IV. DISCUSSION
`The emerging development of compact, low power, small
`size, light weight, and unobtrusive wearable devices can
`facilitate
`remote
`noninvasive monitoring
`of
`vital
`
`physiological signs. Wireless physiological information can
`be useful to monitor soldiers during training exercises and
`combat missions, and help emergency first-responders
`operating in harsh and hazardous environments. Similarly,
`wearable physiological devices could become critical in
`helping to save lives following a civilian mass casualty. The
`primary goal of such a wireless mobile platform would be to
`keep track of an injured person’s vital signs via a short-range
`wirelessly-linked personal area network,
`thus readily
`allowing RF telemetry of vital physiological information to
`command units and remote off-site base stations for
`continuous real-time monitoring by medical experts.
`The preliminary bench testing plotted in Fig. 4 showed
`that the SpO2 and HR readings are within an acceptable
`clinical range. Similarly, the transient changes measured
`during the two breath holding maneuvers confirmed that the
`response time of the custom pulse oximeter is adequate for
`detecting hypoxic episodes.
`
`V. CONCLUSION
`A wireless, wearable, reflectance pulse oximeter has been
`developed based on a small forehead-mounted sensor. The
`battery-operated device employs a
`lightweight optical
`reflectance sensor and incorporates an annular photodetector
`to reduce power consumption. The system has short range
`wireless communication capabilities to transfer SpO2, HR,
`body acceleration, and posture information to a PDA carried
`by medics or first responders. The information could
`enhance the ability of first responders to extend more
`effective medical care, thereby saving the lives of critically
`injured persons.
`
`ACKNOWLEDGMENT
`The authors would like to acknowledge the financial support
`provided by the U.S. Army Medical Research and Material
`Command referenced.
`
`REFERENCES
`[1] G. S. F. Ling, B. K. Day, P. Rhee, and J. M. Ecklund, “In search of
`technological solutions
`to battlefield management of combat
`casualties,” SPIE Conference on Battlefield Biomedical Technologies,
`SPIE vol. 3712, Apr. 1999.
`[2] D. Malan, T. Fulford-Jones, M. Welsh, and S. Moulton, “CodeBlue:
`An ad-hoc sensor network infrastructure for emergency medical care,”
`International Workshop on Wearable and Implantable Body Sensor
`Networks, 2004.
`[3] U. Anliker et al., “AMON: A wearable multiparameter medical
`monitoring and alert system,” IEEE Trans. on Information Technology
`in Biomedicine, 8(4), 2004.
`[4] Y. Mendelson and V. Floroff, “A PDA based ad-hoc mobile wireless
`pulse oximeter,” Proc. IASTED International Conference Telehealth
`2005, Banff, Canada, 2005.
`[5] Y. Mendelson and C. J. Pujary, “Minimization of LED power
`consumption in the design of a wearable pulse oximeter,” IASTED
`International Conference BioMED 2003, Salzburg, Austria, 2003.
`[6] Y. Mendelson and C. J. Pujary, “Measurement site and photodetector
`size considerations in optimizing power consumption of a wearable
`reflectance pulse oximeter,” Proc. of the 25th Annual International
`IEEE/EMBS Conference, Cancun, Mexico, 2003.
`[7] P. Branche and Y. Mendelson, “Signal quality and power
`consumption of a new prototype reflectance pulse oximeter sensor,”
`Proc. of the 31th Annual Northeast Bioengineering Conference,
`Hoboken, NJ, 2005.
`
`915
`
`4
`
`