`
`978-1-4244-6498-2/10/$26.00 ©2010 IEEE
`
`1897
`
`
`Authorized licensed use limited to: Fish & Richardson PC. Downloaded on July 21,2023 at 03:36:19 UTC from IEEE Xplore. Restrictions apply.
`
`1
`
`Implementation of a Wireless Pulse
`Oximeter Based on Wrist Band Sensor
`
`Qing Cai, Jinming Sun, Ling Xia, Xingqun Zhao*
`Collage of Biological Science and Medical Engineering
`Southeast University, Nanjing, China
`*Corresponding author E-mail address: ndt@seu.edu.cn
`
`II.
`
`THE PRINCIPLE OF RELECTANCE OXIMETRY BASED ON
`DIFFUSION PROPAGATION THEORY
`Human tissue is a strongly scattering media, in which the
`movement of light is random. In transmittance oximetry, the
`light detector mainly receives the forward scattered light; while
`in reflectance oximetry, the light source and the detector are
`both placed on the same side of the measurement site, the
`detector mainly receives scattered light after a parabolic path in
`tissue, as shown in figure 1.
`
`Figure 1. Boundary of the parabolic region for photon path distribution
`
`A. Detected light intensity
`For most human tissue, the scattering coefficient is much
`greater than the absorption coefficient. Therefore, when the
`scattering length of tissue is much smaller than its geometric
`dimensions, the photon will experience repeatedly scattering
`before it’s absorbed or escaping outside the tissue boundary. In
`this case, the photon migration in biological tissue can be
`considered as a diffusion process. Light propagation in
`homogeneous biological tissue can be described by photon
`diffusion equation [4]:
`
`Abstract—In order to develop a wrist band pulse oximeter, the
`detected light intensities and effective transmission depth of
`reflectance oximetry are investigated by analyzing the light
`propagation through tissue. A reflectance probe is fabricated
`based on
`the analytical results. An oxygen
`saturation
`measurement system is also implemented and some preliminary
`experiments are conducted. The experimental results have
`confirmed the feasibility of this wrist-band oximeter. Home
`healthcare can be ensured by
`transmitting
`the realtime
`measurements through wireless technology to the community
`monitoring system.
`Keywords- Wrist-band sensor; Oxygen saturation; Remote
`healthcare
`
`INTRODUCTION
`I.
`Non-invasive pulse oximetry has been widely used in
`clinical care for critical ill patient, monitor of patient during the
`anesthesia in operations, research of breath state while sleeping,
`etc. It’s a safe, reliable, real time and continuous measurement
`method. Nowadays transmittance oximetry has been fully
`developed and there’re many mature products which are widely
`and frequently used in hospitals. However, a transmittance
`pulse oximeter is reliable and practical only in case the patient
`is hospitalized or
`lying steadily. Besides,
`transmittance
`measurements sometimes cause discomfort, pain or injury
`because its sensor lightly presses a finger or ear lobe and
`occasionally blocks the blood flow. A reflectance oximeter
`which increases the flexibility of installation can avoid such
`problems and meet the needs of modern home healthcare,
`whereas the research of reflectance oximeters is still in a
`shortage relative to the transmittance type and there are few
`complete solutions for reflectance oximetry[1][2].
`In this study, the detected light intensities and effective
`transmission depth of reflectance oximetry are investigated by
`analyzing light propagation through tissue using the photon
`diffusion equation. We fabricated a reflectance probe based on
`the analytical results, implemented an oxygen saturation
`measurement system and conducted some preliminary
`experiments, whose results have comfirmed the feasibility of
`our oximeter. We also implemented a RF transmitting module
`to found a communication between the measurement system
`and community care center to ensure home monitor of oxygen
`saturation, which is mainly used in night care of patients with
`sleep apnea syndrome [3].
`
`(1)
`
`
`
`φ
`− ∇
`
`( , )r t D
`
`φ
`
`( , )r t
`2
`
`+
`
`μφ
`
`( , )r t
`a
`
`=S
`
`
`( , )r t
`
`∂
`1
`∂
`c t
`( , )r tφ
`Where
` is the fluence rate at the position r in the
`tissue, which means the effective photon density at the position
`aμ
`r at time t. D represents diffusion coefficient,
`is absorption
`coefficient and ( , )r tS
`represents the light source.
`The diffusion coefficient D for photon migration is:
`
`APPLE 1053
`Apple v. Masimo
`IPR2022-01291
`
`
`
`1898
`
`
`Authorized licensed use limited to: Fish & Richardson PC. Downloaded on July 21,2023 at 03:36:19 UTC from IEEE Xplore. Restrictions apply.
`
`=
`
`D
`
`1
`+ −
`(1
`
`μ
`3{
`a
`
`g
`
`μ
`) }
`s
`
`(2)
`
`wave is detected. The distribution function of photon path in
`tissue is [5]:
`
`Thus the density of the photon leaving the tissue is given by
`the flux of the fluence and the detected light intensity is:
`
`
`
`( ,P x y z, )
`
`
`
`2
`
`z
`
`exp(
`
`=
`
`
`
`*[ (k x
`
`2
`
`+
`
`+
`+
`−
`z
`y
`{(k x
`
`)
`2 1 2
`2
`2
`+
`+
`d
`z
`x
`y
`[(
`)
`(
`2 3 2
`2
`2
`+
`+
`d
`y
`z
`1]{[(
`)
`2
`2 1 2
`
`+
`−
`−
`
`d
`[(
`x
`)
`2
`x
`)
`2
`
`−
`+
`+
`
`)
`2
`
`x
`y
`y
`2
`
`+
`+
`y
`2
`2
`+
`z
`]
`2 3 2
`+
`z
`]
`2 1 2
`
`z
`
`] })
`2 1 2
`
`+
`
`1}
`
`(7)
`
`φρ
`t
`( , )
`
`=
`
`π
`Dc
`(4
`)
`
`−
`3 2
`
`−
`5 2
`
`z t
`0
`
`−
`μ
`ct
`exp(
`a
`
`−
`)exp(
`
`+
`z
`d
`2
`2
`0
`Dct
`4
`
`)
`
` (3)
`
`0z is
`
`−
`
`[(1
`
`
`
`)g μ −
`]s
`1
`
`, d is the source-detector
`
`As is shown in figure 1, the photon path in tissue is
`parabolic in the x-z cross section. Under conditions of weak
`absorption (kd<<1), when coordinate is (x,0,0) the z position z0
`(x) of the cambered region is:
`
`2
`
`2
`
`2
`
`2
`
`−
`
`2
`
`−
`
`(8)
`
`2
`
`Where
`separation.
`When time is long enough, the change rate of reflected light
`intensity is as below:
`
`
`
`( )z x
`0
`
`≈
`
`1
`8
`
`⎡
`⎢
`⎣
`
`x
`
`⎡
`⎣
`
`+
`
`(
`
`d
`
`−
`
`x
`
`)
`
`⎤
`⎦
`
`+
`
`−
`
`32 (x d x
`2
`
`)
`
`x
`
`−
`d x
`
`)
`
`(
`
`⎤
`⎥
`⎦
`
`∂
`φρ
`
`, )t
`lim ln (
`∂
`t
`→∞
`t
`
`= −
`
`μ
`c
`a
`
`The maximum depth the incident light reaches is called
`effective transmission depth. z0 reaches the maximum when
`x=d/2;
`
` (4)
`
`z
`
`max
`0
`
`≈
`
`2
`
`d
`
`4
`
`
`
`(9)
`
`This shows that the effective transmission depth depends on
`the source-detector separation d. So choose a suitable distance
`between the light source and detector can ensure the effective
`detection of blood oxygen saturation.
`
`III.
`
`SYSTEM DESIGN AND IMPLEMENTATION
`
`A. System design
`The system consists of wristband sensor, signal acquisition
`and processing module, wireless transmission module, power
`module, display module, etc. Figure 2 shows the system block
`diagram.
`
`Therefore the change rate of reflected light intensity is
`proportional to tissue’s absorption coefficient [4].
`In the near infrared region, the absorption caused by such
`substance as water, cytopigment, etc. is much smaller than that
`caused by oxygenated hemoglobin (HbO2) and reduced
`hemoglobin (Hb), and the light absorbance of HbO2 is different
`from that of Hb at two different wavelengths. Therefore when
`illuminating the tissue containing an arterial bed with light of
`different wavelengths, we can deduce the oxygen saturation
`from the ratio of change rates of reflected light intensity in
`these two wavelengths.
`Note the change rate of reflectance light intensity W, the
`blood oxygen saturation can be expressed as:
`
`(5)
`
`= − ⋅
`SpO A B
`2
`
`W
`W
`
`λ λ
`
`1
`
`2
`
`The light attenuated by body tissue consists of a direct
`current (DC) component and an alternating current (AC)
`component. The DC component is the result of the absorption
`by the body tissue and veins, while the AC component is the
`result of the absorption by the fluctuating volume of arterial
`blood. The change rate of reflectance light intensity can be
`=
`W I
`I
`expressed as:
`, accordingly the formula for
`AC
`DC
`determining oxygen saturation can be obtained :
`
`SpO
`2
`
`= − ⋅
`A B
`
`I
`I
`
`λ
`1
`AC
`λ
`2
`AC
`
`I
`I
`
`λ
`1
`DC
`λ
`2
`DC
`
` (6)
`
`B. Effective transmission depth
`In order to measure oxygen saturation, it is important to
`know whether or not the light reaches the arteries and the pulse
`
`Figure 2. System block diagram
`
`B. Sensor
`A wrist located reflectance sensor is used to sample the
`pulse wave signal. It has functions of optical source control,
`signal conversion and signal amplification. It’s composed of
`light source, light detector, and a wrist band which is used for
`mounting the sensor.
`For reflectance oximetry, the ratio of absorption coefficient
`in the wavelength of 660nm and 900nm is the most sensitive to
`the change in oxygen saturation. To avoid the difference of
`
`2
`
`
`
`1899
`
`
`Authorized licensed use limited to: Fish & Richardson PC. Downloaded on July 21,2023 at 03:36:19 UTC from IEEE Xplore. Restrictions apply.
`
`two
`the distance between
`transmission path caused by
`separated light sources, the system employs dual-wavelength
`light-emitting diode as the light source, with the peak emission
`wavelength of 940nm and 660nm.
`
`formula (10) is implemented for removing the noise, with
`cutoff frequency of 6Hz and -50dB attenuation in 50 Hz and
`above. Where x[n] and y[n] represents the input and output
`signal respectively, while N represents the length of the FIR
`filter. h[n] is the FIR filter coefficients.
`
`N
`
`
`
`[ ]y n
`
`=
`
`
`
`[ ]h n
`
`=
`
`∑
`=
`k
`0
`−
`
`[h N n
`
`−
`[ ] [h k x n k
`
`
`
`]
`
`]
`
`(10)
`
`Figure 3. Wristband sensor
`
`By the formula (3) and (9), both the detected light intensity
`and effective transmission depth are affected by the source-
`detector separation. To ensure sufficient detected light intensity
`and
`transmission depth,
`the optimum
`source-detector
`separation was chose as 8mm based on several experiments.
`The wristband sensor design is shown in Figure 3. To get the
`best measurement effect, the wrist band mounts the sensor on
`the inner wrist right above the arterial vessel.
`
`C. Signal sampling and processing
`To meet the design requirements of miniaturization,
`portable and low power consumption, the system employs
`MSP430 microcontroller as the core of hardware control, data
`processing and transmission.
`The internal 12-bit DAC0 in the MCU periodically send
`two pulses to light source drive module, making light emitting
`diode emit red and infrared light alternatively to obtain optical
`modulation. When the pulse beats, the volume of blood flow
`through the wrist artery changes, the detected red and infrared
`light intensity changed consequently. The ADC sampled the
`output signal of the light detector so as to get the modulated
`photoplethysmography signal.
`
`Figure 4. Signal procesing diagram
`
`The AC component of this signal is effectively extracted
`and amplified through differential amplification performed by
`the built-in operational amplifier OA1 of the microcontroller.
`Figure 4 shows the processing procedure of pulse signal.
`The initial pulse wave signal extracted at this stage
`comprises electromagnetic interference and power supply noise;
`therefore, it’s necessary to eliminate ambient noises in 50Hz
`and above, considering the fundamental frequency of the pulse
`wave is about 1Hz[6]. A symmetric FIR filter as is shown in
`
`Figure 5 shows the initial pulse wave signal extracted by
`the built-in amplifier and effective pulse wave signal after
`filtering. As can be seen from the figure, the system efficiently
`detected the human pulse wave signal, indicating that the
`incident light can visit the arterial vessel.
`
`original signal
`
`1000
`
`900
`
`800
`
`700
`
`600
`
`0
`
`900
`
`800
`
`700
`
`600
`
`0
`
`50
`
`100
`
`150
`
`200
`
`250
`
`300
`
`350
`
`400
`
`450
`
`500
`
`Red
`
`digitally filtered signal
`
`50
`
`100
`
`150
`
`200
`
`250
`
`300
`
`350
`
`400
`
`450
`
`500
`
`Red
`
`Figure 5. Original signal(upper) and filtered signal(lower) of pulse wave
`
`The heart rate is measured by calculating the cycle of pulse
`wave signal, and the oxygen saturation is deduced by
`computing the ratio of the light intensity change rates under
`two different wavelengths. The real time measurement results
`and pulse wave are displayed on an OLED continuously
`
`D. Wireless data transmission
`The wireless transmission module founds a communication
`between measurement system and community care system,
`transfers the measurement results to a PC. Therefore, a number
`of users can be centralized monitored by creating a small
`wireless network.
`The wireless transmitting module consists of a transceiver
`CC2500, a chip antenna, etc. CC2500 is a single chip
`transceiver designed for very low power wireless applications
`and intended for the ISM frequency band at 2.4 GHz. Wireless
`receiving module integrates a MSP430 MCU, CC2500 RF
`transceiver and a USB interface. It can be directly connected to
`the PC of community care center.
`Wireless transmitting module communicates data with
`measurement system through the SPI interface and sends
`measurement results
`to
`the receiving module. Wireless
`receiving module transfers the received data to the PC of
`community care system via USB interface. RF chip CC2500
`keeps control of RF module and IEEE 802.15.4 protocol to
`regulate the wireless transmission of measurement data.
`
`3
`
`
`
`1900
`
`
`Authorized licensed use limited to: Fish & Richardson PC. Downloaded on July 21,2023 at 03:36:19 UTC from IEEE Xplore. Restrictions apply.
`
`Several oximeters can easily set up a star wireless network [7].
`Transmission distance can reach 100m in an open space, relay
`stations are needed while transferring data for farther distance
`or through walls.
`
`IV. EXPERIMENTAL RESULTS
`To test the reliability of this system, an apnea experiments
`is conducted; five healthy people took part in this experiment.
`We recorded the values of R at time intervals during the slowly
`recovery process after several seconds of breath holding, and
`drew the value changes in line graphs as shown in Figure 6.
`
`continuously and non-invasively through a wearable reflection
`sensor placed on the inner wrist; its facility makes it quite
`suitable for daily home care. From the waveform and
`measurement results it can be seen that the system can achieve
`the wrist pulse signal acquisition and noise filtering. The
`preliminary experiments have confirmed the reliability of the
`measurement results. The wireless
`transmission module
`transfers real time measurement data to community care center,
`providing the basis for community centralized care and remote
`monitoring. This wrist band pulse oximeter offers an effective
`means of home monitoring oxygen saturation, which has a
`broad application scope and aspects
`
`REFERENCES
`[1] Miia M, Antti K, Esko A, YoungChung W. Optimum Place for
`Measuring Pulse Oximeter Signal in Wireless Sensor-Belt or Wrist-Band.
`International Conference on Convergence Information Technology,
`2007 .
`[2] O. Abdallah, A. Bolz. Signal Processing by Reflectance Pulse Oximetry
`for Monitoring the fractional Oxygen Saturation and the Detektion of
`Anemia. IFMBE Proceedings, 2(7): 1152-1155.
`[3] Chaudhary B,Dasti S,Park Y,et a1.Hour-to-hour variability of
`oxygen saturation in sleep apnea.Chest,1998,113(3):719-722.
`[4] Y. Wei, G. Qingen, Y. Xiang, L. Feixue, D. QinKai. ExperimentaI study
`on non-invasive detection of refIectance oxygen saturation. Chinese
`Medical Equipment Journal, 2007;28(10): 4-7.
`[5] ShechaoFeng, Fan-AnZeng and Britton Chance. Photon migration in the
`presence of a single defect: a pertubation analysis. Appl.Opt.,1995,
`34(19): 3826-3837.
`[6] Frans M .Coetzee and Ziad Elghazzawi. Noise-Resistant Pulse Oximetry
`Using a Synthetic Reference Signal,IEEE Transaction on Biomedical
`Engineering,2000;47(8): 1018-1026.
`[7] M. Site, L. Jun. Design of Wireless Network for Multi- par ameter
`Patient Monitor Based on ZigBee Technology. Chinese Medical
`Equipment Journal, 2008;29(9): 53-56.
`
`Figure 6. Change map of the value of R after breath holding for 10s and 25s
`
`About 6s after the apnea, R values began to rise to a peak,
`which means oxygen saturation began to decrease. It can be
`seen from the comparison of these two graph that for the longer
`time of breath holding, the increasing amplitude of R is larger,
`that is the decreasing extend of oxygen saturation is larger, and
`the recovery time is longer as well.
`The experimental results indicate that this system can
`effectively detect the change of oxygen saturation
`
`V. CONCLUSION
`This paper describes the system design of a wristband pulse
`oximer and completes the prototype fabrication. This oximeter
`can obtain oxygen saturation and heart rate information
`
`4
`
`