CAPACITIVE SENSING FOR PULSE RATE MONITORING

R. Niall Tait

Dept. of Electronics, Carleton University, 1125 Colonel By Drive, Ottawa, Ontario, K1S5B6, Canada

Keywords: Capacitance, Sensor, Pulse.

Abstract: This paper describes pulse rate monitoring using capacitive electrodes on finger or wrist. The technique

uses a single frequency measurement suitable for low cost and low power applications. The system may

enable convenient, comfortable, and continuous monitoring. Pulsatile flow resulted in approximately 5-10

fF of capacitance variation, a level easily measured using inexpensive capacitance measurement integrated

circuits.

1 INTRODUCTION

Automated pulse and heart rate measurements have

become commonplace with the development of

personal training and health monitoring devices.

These devices are accurate and easy to use.

However ECG based systems require either a chest

strap which may be uncomfortable or inconvenient,

or require use of a finger touch sensor and therefore

do not provide continuous monitoring. Pressure

based systems require a finger, wrist, or arm cuff

that periodically inflates, can be uncomfortable, and

does not provide continuous monitoring. Optical

systems (pulse oximeters) require an optical

transmitter with adequate intensity output to

penetrate tissue, and although many battery operated

models are available, they are not compatible with

very low power operation.

Bioimpedance measurement methods have found

many applications, however they have seldom been

applied to simple pulse rate monitoring. One reason

may be that obtaining the wealth of information

available from a complex impedance measurement

requires at least four contact electrodes usually using

gel on at least one limb, wired to precision bench top

electrical instrumentation (Bayford 2006). However

if capacitance is the only parameter to be measured,

the situation is changed.

Capacitance sensing has become very common in

recent years. Traditionally capacitance sensing

suffered from a reputation of being susceptible to

parasitic effects and requiring unstable high input

impedance circuitry. Most measurements involved

bridge circuits (Mohanty 2004) or network analyzers

(Ferrier 2008). However those complexities have

largely been overcome using microcontroller based

capacitance measuring circuits. These sensors are

now widely used in consumer electronics for non-

contact switches, sliders, and track pads. In addition

to the benefit that capacitance switches do not

require direct electrical contact, they also draw no

direct current in any state, and are therefore suitable

for low power applications. With no direct current

the devices are inherently low noise, although

support circuitry may not share this benefit.

These now ubiquitous capacitance sensing IC’s

have been demonstrated in many applications

including chromatography systems (Takeuchi

2008). They also have potential to be used in finger

or wrist band pulse rate monitors, enabling low cost

and very low power operation for ambulatory

monitoring, or in combination with other sensors for

more complete vital sign monitoring.

2 METHODS

2.1 Experimental

In order to test the feasibility of measuring pulse rate

using finger capacitance measurement, the Analog

Devices AD7746 capacitance to digital converter

was used. This integrated circuit is specified to

provide 4fF accuracy and 4 aF resolution at a 32kHz

measurement frequency, and has a 2 wire I

compatible digital interface (Analog Devices, Inc.

2005). The IC is available as part of an evaluation

board, using a Cypress Microsystems CY7C68013

215

Niall Tait R. (2010).

CAPACITIVE SENSING FOR PULSE RATE MONITORING.

In Proceedings of the Third International Conference on Biomedical Electronics and Devices, pages 215-218

DOI: 10.5220/0002750902150218

 SciTePress

Ez-USB microcontroller for the I

C to USB interface

to a personal computer for programming and data

acquisition (Cypress Semiconductor Corporation

2005).

A schematic of the measurement system is

shown in Figure 1. A key factor enabling the

measurement is close contact between the finger and

electrode. This can be achieved by incorporating

electrodes into a finger clip similar to a pulse

oximeter, or incorporating them into a cuff using a

fabric hook-and-loop closure. The cuff approach

can be less bulky and more secure than the clip,

although the cables may be inconvenient.

Figure 1: Schematic of finger pulse measurement.

Excitation and measurement electrodes were

constructed from copper tape strips 6 mm wide and

14 mm long covered with dielectric tape. Electrodes

were adhesive mounted on a fabric hook-and-loop

cuff for attachment to the finger. Electrodes were

connected to the AD7746 using coaxial cables

approximately 400 mm long. Measurements were

completed with the subject’s arm resting on a table

top approximately 20 cm below the level of the

heart. The capacitance measurement circuit was set

up using the evaluation board software (Analog

Devices, Inc. 2005). The IC was set to a typical

configuration with one excitation electrode driven

by excitation channel B with an amplitude of V

/2,

and one input electrode connected to the positive

terminal of input channel 1. The capacitance

measurement was then single-ended with continuous

sampling at a rate of 16.1 Hz. No further signal

processing was used.

For wrist measurements, electrodes with

dimensions identical to those used in the finger

monitor were located on either side of the radial

artery. These were also fixed in place using a fabric

hook-and-loop band. The measurement procedure

was identical to the finger measurement.

This work was intended only as an initial

feasibility study for this technique and no efforts

were made to evaluate the effect of position and

motion artifacts, or the variation in results for

different subjects.

2.2 Electrical Model

An electrical model of the pulsatile flow is required

in order to enable discussion of the measured results

with respect to the physical system. At this stage a

simple model of the finger tissue can be used,

replacing it with a parallel capacitance and

resistance. However it is important to include series

capacitance between the finger and electrodes

representing the dielectric and any small air gap.

The model shown in Figure 2 can be easily

evaluated using a circuit simulator, numerical

calculation, or analytical methods.

Figure 2: Electrical model for finger impedance.

Permittivity of the finger tissue is high, and

depends on the exact composition of the tissue.

However if the relative permittivity is approximately

=3000 and the finger diameter is 13 mm, the finger

capacitance using a parallel plate approximation

should be about 170 pF.

The signal to be measured is associated with the

pulsatile nature of the blood flow. The finger is not

a rigid structure, so the increase in blood pressure

during systole should correspond to an increase in

volume of the finger. In addition, if the electrodes

are attached to a rigid band, there will be a

corresponding decrease in the gap between the

finger and the electrodes.

A perceived drawback of a capacitive

measurement is that the signal will be loaded by

both the shunt resistance and the series capacitance

components. The finger resistance will influence the

measured value as it passes current, but provided the

resistance is fairly high and the measurement

frequency is high enough the effect is small. The

dielectric coating on the electrodes and any air gap

between electrodes and skin represent a low

permittivity and small capacitance. However due to

the series connection, these capacitances dominate

the overall impedance of the system regardless of

the measurement frequency. Fortunately, for a pulse

measurement the exact amplitude of the measured

impedance is much less important than the

BIODEVICES 2010 - International Conference on Biomedical Electronics and Devices

216

frequency provided the amplitude is adequate for

measurement.

For a high resistance subject, the measured

capacitance then is expected to be the series

combination of the finger capacitance and the two

electrode capacitances. Using the dimensions for

electrode and finger mentioned above, and an air gap

of 0.5 mm between electrode and skin surface

produces a total capacitance of 0.74 pF. Further

adding a sinusoidal variation of 1% in finger

diameter (and corresponding reduction in gap) a

capacitance fluctuation of about 8 fF is predicted.

3 RESULTS AND DISCUSSION

Initial experiments were completed using electrodes

on the index finger. The capacitance variation with

time is shown in Figure 3. It corresponds well with

the pulse and shows features similar to a typical

infrared absorption measurement. No attempt was

made to shield electrodes in this measurement, and

the capacitance values are sensitive to parasitic

coupling to nearby objects and to movement of the

subject. These effects require further study, but are

commonly encountered and can be reduced through

improved design and signal processing (Kim 2006).

Figure 3: Capacitance measurement for electrodes

mounted on the finger.

The mean capacitance shown in Figure 3 is 0.565

pF with a fluctuation around 8 fF, which is similar to

the estimate from the electrical equivalent model.

Detailed noise measurements were not completed,

but the RMS noise level in stable operation and with

no finger between the electrodes is less than 100 aF.

The signal was not observed to show significant

variation due to position on the finger or tension in

the cuff. No quantitative measurements were made

to investigate the effect of changing conductivity on

the capacitance measurement.

A second measurement location was investigated

using parallel electrodes held on either side of the

radial artery using a wrist band. With this electrode

arrangement the capacitance is between the facing

edges of the electrodes, and the change is expected

to be lower than with the finger measurement. In

fact the amplitude variation in the capacitance was

found to be similar to the finger measurement,

however the total capacitance and noise level was

higher as shown in Figure 4.

Figure 4: Capacitance measurement for electrodes

mounted on the wrist.

While the pulse can be identified in this signal, an

improvement in the signal amplitude or noise is

required to make this a practical measurement. If

this can be improved the implementation would be

much more convenient for a subject using the sensor

for continuous pulse monitoring as it could be

incorporated in a wristwatch style instrument.

However placement of the electrodes relative to the

radial artery will be much more challenging than

placement of the finger-cuff electrodes.

4 CONCLUSIONS

Capacitive measurement of pulse rate has been

demonstrated to be an interesting technology for low

cost and low power applications. The most

promising results have been found using a finger

cuff containing measurement electrodes which

provided a pulsatile variation of approximately 8 fF.

ACKNOWLEDGEMENTS

This work was supported by the Natural Sciences

and Engineering Research Council of Canada

(NSERC).

REFERENCES

Analog Devices, Inc. AD7745/AD7746 24-Bit

Capacitance-to-Digital Converter with Temperature

Sensor. Norwood, MA, USA: www.analog.com, 2005.

Analog Devices, Inc. AD7746 Evaluation Board.

Norwood, MA, USA: www.analog.com, 2005.

Bayford, R.H. “Bioimpedance Tomography.” Annu. Rev.

Biomed. Eng., 2006: 63-91.

2.095

2.100

2.105

2.110

Time(s)

0.560

0.564

0.568

Time(s) 6

0.572

2.115

CAPACITIVE SENSING FOR PULSE RATE MONITORING

217

Cypress Semiconductor Corporation. CY7C68013 EZ-USB

FX2 USB Microcontroller. San Jose, CA, USA:

www.cypress.com, 2005.

Ferrier, G.A., Hladio, A.N, Thomson, D.J., Bridges, G.E.,

Hedayatipoor, M., Olson, S., Freeman , M.R.

“Microfluidic Electromanipulation with Capacitive

Detection for the Mechanical Analysis of Cells.”

Biomicrofluidics, 2008: 044102.

Kim, B.S., Yoo, S.K. “Motion artifact reduction in

photoplethysmography using independent component

analysis.” IEEE Transactions on Biomedical

Engineering 53, no. 3 (2006): 566-568.

Mohanty, S.K., Sohn, L.L., Beebe, D.J. “Hybrid

Polymer/Thin Film Impedance System for Label-free

Monitoring of Cells.” EMBS. San Francisco: IEEE,

2004. 2561-2565.

Takeuchi, M., Li, Q., Yang, B., Dasgupta, P.K., Wilde,

V.E. “Use of a Capacitance Measurement Device for

Surrogate Noncontact Conductance Measurement.”

Talanta, 2008: 617-620.

BIODEVICES 2010 - International Conference on Biomedical Electronics and Devices

218