A Wireless Device for Measuring Hand-Applied Forces

8/9/2019 A Wireless Device for Measuring Hand-Applied Forces

http://slidepdf.com/reader/full/a-wireless-device-for-measuring-hand-applied-forces 1/4

Abstract — We report on a wireless, electromyography(EMG)-based, force-measuring system developed to quantifyhand-applied loads without interfering with grasping function.A portable surface EMG device detects and converts to voltageoutput biopotentials generated by muscle contractions in theforearm and upper arm during hand-gripping and tractionactivities. After amplifying and bandpass filtering, our radiofrequency (RF)-based design operating at ~916 MHz wirelesslytransmits those voltages to a data acquisition (DAQ) system upto 20 meters away. A separate calibration system is used torelate an individual user’s EMG signal to known pull andclenching forces during specific applications. Real-time EMGdata is processed and displayed in software developed withLabView™ (National Instruments, Austin, TX). Data is then

converted to force data using individual calibration curves.With EMG electrodes placed over any major forearmmuscle, calibration curves for seven subjects demonstratedlinearity (R 2 > 0.9) and repeatability (<10% of average slope) to110 newtons (N). Preliminary results in clinical application onnewborn delivery suggest that this approach may be effectivein providing an unobtrusive and accurate method of measuringhand-applied forces in applications such as rehabilitation andtraining.

Keywords — delivery, electromyography, EMG, handforces, newborn, obstetrics, rehabilitation, tactile sensing,wireless

I. I NTRODUCTION

Applications for measuring hand-applied forces havevaried from basic biomechanics research [1] to medical

procedures such as surgery [2] and birth [3] to rehabilitation[4]. Methods include integrated electromyography (EMG),[5]-[7] modifying commercially available polymer sensingelements, [8] and designing and packing a silicon diaphragm[9]. In addition, wireless communication has been used inclinical applications including measurement of blood

pressure, oxygen saturation and orthopedic device loading[10].

For some applications, a method by which to measureforce without interfering with the grasping process isdesired. This paper reports a method to indirectly measure

hand-applied forces without interfering with the grasping process.

II. S YTEM DESIGN A ND CONSTRUCTION

Our system for measuring hand-applied forces consistsof a wireless, EMG device and a DAQ system. There aretwo sub-systems to the wireless EMG device: the portableEMG unit (Fig. 1a) and the receiver unit (Fig. 1b). The

portable unit collects the recorded EMG signal and transmits

it to the receiver unit and computer (for data display, storageand post-processing). The EMG signal is collected throughsurface electrodes positioned on the skin overlying aspecific muscle of interest; EMG leads for the electrodes areattached to the transmitter unit.

Fig. 1. Wireless EMG Device in Two Parts: (a) the transmitter unit detects the EMG signal from the EMG leads, which is thenamplified by an instrumentation amplifier. To improve thedevice’s signal-to-noise ratio, the gain of the instrumentationamplifier can be adjusted with the help of the LED indicators.(b) The receiver unit processes the transmitted signal, which isthen imported by the data acquisition hardware and software.

The EMG signal initially passes through aninstrumentation amplifier, which takes the weak electricsignal (0.1 – 5mV) [11] from the muscle and increases itsamplitude to above 0.5V so that it can be further processedand recorded. The instrumentation amplifier (Burr-BrownINA128P) is connected with a 505-ohm variable resistor sothat amplification gain can be adjusted for each user. Anideal gain for most clinical uses ranges from 100X to10000X. There are two light-emitting diodes (LED’s)connected to comparators, which are designed to signal gainsaturation.

Once amplified, the biopotential output passes through a bandpass filter that isolates frequencies between 0.1 kHzand 1 kHz; this improves the signal-to-noise ratio (SNR).The filtered EMG signal then passes through the wirelesstransmission unit.

The transmitter module (Linx TXM-916-ES) andtransmitter antenna (Linx ANT-916-SP) are located in the

portable unit, while the receiver module (Linx RXM-916-ES) and receiver antenna (Linx ANT-916-SP) are located inthe receiver unit connected to the DAQ system. The wirelessmodules, chosen for their compatibility with EMGfrequency ranges [11], operate at a frequency of 916.5 MHzand are capable of transmitting analog signals of 20 Hz to 28kHz up to a distance of 20 meters. To avoid interference,the frequency at which the wireless modules operate ishigher than the two medical frequency bands, 174-216 MHzand 450-470 MHz, prescribed by the Federal

A Wireless Device for Measuring Hand-Applied Forces

William Tam 1, Robert H. Allen 1, Yen Shi Gillian Hoe 1, Stanley Huang 1,I-Jean Khoo 1, Katherine E. Outland 1 and Edith D. Gurewitsch 2 Departments of 1Biomedical Engineering and 2Gynecology and Obstetrics

Johns Hopkins University, Baltimore, Maryland

(a) Transmitter Unit(b) Receiver

Unit

Authorized licensed use limited to: SARVAJANIK COLLEGE OF ENG & TECH. Downloaded on September 14, 2009 at 05:17 from IEEE Xplore. Restrictions apply.

0-7803-8439-3/04/$20.00©2004 IEEE 2121

Proceedings of the 26th Annual International Conference of the IEEE EMBS San Francisco, CA, USA • September 1-5, 2004



Communications Commission (FCC), and higher than thetelevision band frequency, 470-862 MHz [10]. Thus, our system does not interfere with wireless transmissions fromother equipment present in the same clinical environment,

but still operates within a safe range for medicalapplications. Two 9V batteries power both the transmitter and receiver units. With an expected alkaline battery life for the receiver and transmitter units are 29 and 33 hours,respectively.

To ensure preservation of the desired EMG signal withinthe appropriate frequency range and remove noise acquiredduring wireless transmission, the transmitted biopotential

passes through a second bandpass filter with cutoff frequencies between 0.1 kHz and 1 kHz. An operationalamplifier boosts the gain of the EMG signal after wirelessreception. A half-wave rectifier discards the negative

portion of the EMG signal. Finally, the positive EMG signalis integrated, smoothing the spiked EMG signal into a moreeasily comprehensible temporal envelope. This signal thenarrives at the DAQ board for recording and data analysis. Aschematic of the entire wireless EMG system is shown inFig. 2.

Fig. 2. Schematic of Wireless EMG Device.

The DAQ system is specifically designed to digitize and process incoming analog signals. Implemented inLabView™, the program provides for live-data display anddata storage for post-processing. The program also promptsthe user for application-related information, which is linkedwith output files generated by that specific execution. Toensure the system’s portability, the code was implementedon a laptop computer using a 16-bit PCMCIA card for dataacquisition and a terminal block; the hardware is

programmed to acquire data at rates up to 1 kHz.The program guides the user through a manual

calibration process, which involves a series of pulls atdifferent pre-set force ranges. Depending on the application,calibration force values range from 20 N to 110 N. After the

calibration process, the program displays the real-timevoltage from the EMG. The voltage from the wireless EMGis then stored for post-processing.

III . SYSTEM PERFORMANCE A ND R ESULTS

To determine efficacy, we calibrated the system in aseries of laboratory tests and then performed in vivo clinicaltesting after obtaining Institutional Review Board approval.

Most hand-gripping and pulling efforts require contractionof the forearm and upper arm muscles [6], [7]. To determine

placement of electrodes in our single-channel system, weused two calibration systems: a JAMAR™ hand-dynamometer for grip force and a modified Chatillon™dynamometer for traction forces. We tested five forearmand four upper arm muscles, including the flexor carpiradialis, flexor carpi ulnaris, brachioradialis, flexor digitorum profundus, palmaris longus, and extensor indicis

proprius in the forearm and the triceps brachii, biceps brachii, lateral deltoids and lower trapezius muscles in theupper arm. Fig. 3 shows a typical calibration for a tractiontest, where R 2 values ranged between 0.90 and 0.99 for traction tests up to 90 N for the four muscles in the forearm.Fig. 4 shows a typical calibration for a grip test up to 300 Nin two subjects. Similar results (not shown) were obtainedfor other subjects and in tests with the upper arm muscles.For traction force applications in the range of up to 110 N intraction, the largest gradient and voltage output wereobtained from muscle contractions recorded from the

palmaris longus. As this force type and force range were of interest, all further testing was conducted using this muscle.

Fig. 3. Calibration Curve: The results of one subject performingthe calibration process four times in traction show theconsistency of the device as the slopes of the trendlines vary

between 0.0089 – 0.0114. Multiple tests also confirm the linear relationship between EMG voltage and pull force.

y = 0.0011x + 0.91R 2 = 0.99

y = 0.0013x + 0.85R 2 = 0.95

y = 0.0015x + 0.83R 2 = 0.97

y = 0.0018x + 0.89R2 = 0.99

y = 0.0016x + 0.70

R2

= 0.98y = 0.0017x + 0.70

R 2 = 0.980.6

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

1.6

0 50 100 150 200 250 300 350 400

Grasping Force (N)

E M G

V o

l t a g e

( V )

Trial 1 (A)Trial 2 (A)Trial 3 (A)Trial 1 (B)Trial 2 (B)Trial 3 (B)

Fig. 4. Calibration Curve: The results of two subjects (A and B)each performing three calibration trials with the hand-grip test.Consistency is demonstrated here by trendlines and R-squaredvalues.

Seven subjects performed traction tests with theChatillon™ dynamometer. To mimic the downward traction

Authorized licensed use limited to: SARVAJANIK COLLEGE OF ENG & TECH. Downloaded on September 14, 2009 at 05:17 from IEEE Xplore. Restrictions apply.

2122





become impacted behind the maternal pubic bone. Anattending physician was able to step in after about 90seconds and perform maneuvers that minimized traction onthe fetal head, which then resulted in atraumatic delivery.The data for all five deliveries are summarized in Table I;they include RVD, SD, VAVD and cesarean section (C/S).In our clinical testing, the force values we measured with theEMG device during routine, difficult and shoulder dystociadeliveries were similar to those measured using

piezoresistive sensors [3], [13].

TABLE ISUMMARY OF DELIVERIES USING SYSTEM

Trial Clinician DeliveryMode

BirthWeight (g)

DeliveryComplications

Max. Force(N)

1 A RVD 2650 None 37.82 B SD 3477 Severe SD >110 1

3 C VAVD 3570 Difficult 64.24 D SD 3761 Severe SD >110 2

5 C C/S 3487 None 36.91Linear limit exceeded; peak force reading was 146.3N2Linear limit exceeded; peak force reading was 187.1N

IV. S UMMARY A ND CONCLUSIONS

We have developed a wireless and portable force-sensing system that does not interfere with the grasping

process Preliminary laboratory and clinical testing suggeststhat some applications may benefit from this approach toforce measurement during quasi-static applications. Thewireless aspect, although long used in other biomonitoringapplications, has not been previously reported in EMGstudies [10]. The method confirms that integrated EMG isrelated to the force of exertion for a given posture [14] andarm position [7]. Within 0 N to 110 N of applied force, the

palmaris longus muscle generates linear (R 2 > 0.9) andconsistent (<10% of average slope) recordings, althoughtemporal variability requires frequent calibration [5].

While the tactile sensing community has not embracedusing EMG for generalized force measurement because of its dependence on arm position, [6], [8] the results heresuggest that EMG can be used for quasi-static forcemeasurement where the relative arm position is known.However, further testing is needed to confirm the linearityand consistency for a greater range of force measurement.We believe that with the addition of a second EMG channel,the simultaneous recording of two muscle activities can becombined to provide a more clear and accurate picture of hand-applied force. In addition, limits for this system’sapplication need to be quantified; if successful, EMG forcesensing may be useful for other quasi-static, in vivo applications, such as dentistry and rehabilitation.

ACKNOWLEDGMENT

The authors thank Paul Gilka for his technical assistanceduring prototype development. Kiran Sahni, Shoichi Okada,Glen Quigley, and Esther Kim provided invaluableassistance with laboratory and clinical data collection.

R EFERENCES

[1] K. N. An, E. Y. Chao, W. P. Cooney, and R.L. Lincheid. “Forces in thenormal and abnormal hand,” J. Orthop. Res ., vol. 3, pp. 203-211, 1985.

[2] R. A. Dickson, F. V. Nicolle, J. S. Calnan, and A. Petrie, “A device for measuring the force of the digits of the hand,” Biomed. Eng ., vol. 7,1972.

[3] J. Sorab, R. Allen, and B. Gonik, “Tactile sensory monitoring of clinician-applied forces during delivery of newborns,” IEEE Trans.

Biomed. Eng ., vol. 35, pp. 1090-1093, 1988.[4] K. L. Kilgore, P. H. Peckham, G. B. Thrope, M. W. Keith, and K. A.

Gallagher-Stone, “Synthesis of Hand Grasp Using Functional Neuromuscular Stimulation,” IEEE Trans. Biomed. Eng ., vol. 36, pp.761-769, 1989.

[5] T. J. Armstrong, D. B. Chaffin, and J. A. Foulke, “ A methodology for

documenting hand positions and forces during manual work,” J. Biomech ., vol. 12, pp. 131-133, 1979.[6] J. Duque, D. Masset, and J. Malchaire, “Evaluation of handgrip force

from EMG measurements,” Appl. Ergonom ., vol. 26, pp. 61-66, 1995.[7] P. J. Kier, “The effects of posture on forearm muscle loading during

gripping,” Ergonom ., vol. 46, pp. 956-975, 2003.[8] R. T. Jensen, R. G. Radwin, and J. G. Webster, “A conductive polymer

sensor for measuring external finger forces,” J. Biomech ., vol. 24, pp.851-855, 1991.

[9] D. J. Beebe, D. D. Denton, R. G. Radwin, and J. G. Webster, “A silicon- based tactile sensor for finger-mounted applications,” IEEE Trans. Biomed. Eng ., vol. 45, pp. 151-159, 1998.

[10] T. F. Budinger, “Biomonitoring with wireless communications,” Ann. Rev. Biomed. Eng , Vol. 5, pp. 383-312, 2003.

[11] J. G. Webster. Medical Instrumentation: Application and Design , NewYork: John Wiley, 1998.

[12] R. H. Allen, B. R. Bankoski, C. A. Butzin, and D. A. Nagey,“Comparing clinician-applied loads for routine, difficult and shoulder dystocia deliveries,” Am. J. Obste. Gynecol ., vol. 171, pp. 1621-7, 1994.

[13] R. Allen, J. Sorab, and B. Gonik, “Risk factors for shoulder dystocia:An engineering study of clinician-applied forces,” Obstet. Gyneco., vol.77, pp. 352-5, 1991.

[14] S. Bouissett. “EMG and muscle force in normal motor activities,” In: Electromyography and Clinical Neurophysiology and Clinical Neurophysiology , Ed. Desmedt JE. Basil: Kargor, 1973, pp. 547-583.

Authorized licensed use limited to: SARVAJANIK COLLEGE OF ENG & TECH Downloaded on September 14 2009 at 05:17 from IEEE Xplore Restrictions apply

2124

Documents

A Wireless Device for Measuring Hand-Applied Forces