GLOBAL BIOMECHANICAL EVALUATION DURING WORK

AND DAILY-LIFE ACTIVITIES

Francesco Draicchio, Alessio Silvetti, Federica Amici, Sergio Iavicoli and Alberto Ranavolo

Istituto Superiore per la Prevenzione e la Sicurezza del Lavoro (ISPESL), Monteporzio, Italy

Rossana Muscillo, Maurizio Schmid, Tommaso D’Alessio

Dpt. Elettronica Applicata, Università di Roma Tre, Via della Vasca Navale, 84, Roma, Italy

Giorgio Sandrini, Michelangelo Bartolo

Istituto Mondino, Università di Pavia, Via Ferrata Adolfo, 27100 Pavia, Italy

Giancarlo Orengo, Giovanni Saggio, Carmela Conte

Dpt. Ingegneria Elettronica, Università Tor Vergata, via Politecnico 1, 00133 Roma, Italy

Keywords: Gait analysis, Accelerometer, Gyroscope.

Abstract: Advances in technology in the last decades have provided the opportunity to observe human behaviour in

the three dimensional space with great spatial accuracy. Optoelectronic techniques for measurements of

human motions have been developed. However, it is found that, in the work environments, these methods

are complicated to set up and can only easily be applied in laboratory. On the other hand, electronic sensors

such as accelerometers and gyroscopes, have been developed and applied to solve the relevant outdoor

application problems of the image-based methods. These sensors have been evaluated for the 3D

measurement of trunk, lower and upper segments, during posture, walking and rising from a chair, in both

normal and pathological conditions. In the present study we used a device including accelerometers and

gyroscopes in order to calculate the angular behaviour of the pelvis on the sagittal, frontal and horizontal

plane, during the following tasks: walking, gait initiation, gait termination, seat-to-stand and stand-to-seat,

squat, standing anterior and lateral reaching and grasping, anterior and lateral trunk flexion and trunk

rotation. The assessment of pelvis during posture and movement is important in improving our

understanding of the motor strategies at work and preventing injuries (i.e. low back pain) and mechanical

whole body fatigue. The calculated angles were compared to that computed by a high-quality optical motion

analysis system (SMART-E System, BTS, Milan, Italy) consisting of eight infra-red cameras (operating at

120 fps) to detect the movements in three-dimensional space of three retro-reflective markers (15 mm

diameter). For the comparison of the Range of Motions (ROMs) we used the root mean squared error

(RMS) whereas the Coefficient of Multiple Correlation (CMC) was used to evaluate overall waveform

similarity of instantaneous angle curves. Preliminary results showed a high similarity between the extracted

angle tracks (anterior-posterior behaviour on the sagittal plane, pelvic obliquity and intra-extra rotation of

the pelvis) in all of the acquired tasks. We also found low errors in the computation of the corresponding

ROMs. This study suggests to apply an accurate, inexpensive and simple method to measure the kinematics

of the pelvis during common work and daily-life activities.

1 INTRODUCTION

Advances in technology in the last decades have

provided the opportunity to observe human

behaviour in the three dimensional space with great

spatial accuracy. Image-based methods for the

measurement of human motion have been

developed, such as optoelectronic techniques

(Medved, 2001; Cappozzo, 2005). However, it is

108

Draicchio F., Silvetti A., Amici F., Iavicoli S., Ranavolo A., Muscillo R., Schmid M., D’Alessio T., Sandrini G., Bartolo M., Orengo G., Saggio G. and Conte

C. (2010).

GLOBAL BIOMECHANICAL EVALUATION DURING WORK AND DAILY-LIFE ACTIVITIES.

In Proceedings of the Third International Conference on Biomedical Electronics and Devices, pages 108-112

DOI: 10.5220/0002740901080112

 SciTePress

found that, in work environments, these methods are

complicated to set up and can only easily be applied

in laboratory. On the other hand, electronic sensors

able to provide orientation based on accelerometers

and gyroscopes have been developed and applied to

solve the relevant outdoor application problems of

the image-based methods. These sensors have been

used and evaluated for the 3D measurement of trunk,

lower and upper segments, during posture, walking

and rising from a chair, in both normal and

pathological conditions (Pfau, 2005, Lau, 2008,

Plamondon, 2007, Coley, 2007, Veltink, 2007,

Boonstra, 2006, Zijlstra, 2008).

In the present study we compared the angular

behaviour of the pelvis in the sagittal, frontal and

horizontal plane calculated with a wearable inertial

device including triaxial accelerometers and triaxial

gyroscopes with that computed with a high precision

and accuracy optoelectronic motion analysis system.

We recorded the angle trajectories and excursions

during the following tasks: standing anterior

reaching and grasping, standing oblique reaching

and grasping, standing oblique opposite reaching

and grasping, anterior trunk flexion and sit to stand.

The assessment of pelvis during posture and

movement is important in improving our

understanding of the motor strategies at work and

preventing injuries (i.e. low back pain) and

mechanical whole body fatigue.

2 MATERIALS AND METHODS

Ten healthy male subjects (mean age 38 ± 4 years,

range 20-55 years) were enrolled. All gave their

written informed consent after receiving a full

explanation of the study, which conformed to the

requirements of the Declaration of Helsinki.

We used a Wi-Fi transmission miniaturized

device integrating an accelerometer and a gyroscope

(MicroStrain 3DM-GX2, MicroStrain, Inc.,

Williston, USA) placed directly on the skin over the

sacrum. The device offers a range of output data

quantities from fully calibrated inertial

measurements to computed orientation estimates.

All quantities are fully temperature compensated and

corrected for sensor misalignment. The angular rate

quantities are further corrected for G-sensitivity and

scale factor non-linearity to third order.

The extracted curves and the calculated angles

were compared to those simultaneously acquired by

a high-quality optical motion analysis system

(SMART-E System, BTS, Milan, Italy, Ferrigno and

Pedotti 1985) consisting of eight infra-red ray

cameras (operating at 120 fps) to detect the

movements in three-dimensional space of three

retro-reflective markers placed on the skin over the

sacrum and the right and left anterior superior iliac

spinae. Data processing was performed using

Analyzer software (BTS, Milan, Italy).

Before starting formal measurements, all

subjects did a practice session to familiarize

themselves with the experimental procedure and

with the tasks consisting of eleven movements

performed in a quiet room with normal indoor

temperature and lighting. The performed tasks were:

standing anterior reaching and grasping, standing

oblique reaching and grasping, standing oblique

opposite reaching and grasping, anterior trunk

flexion and sit to stand. Standing reaching and

grasping tasks have been performed with the subject

starting from a standing posture, with the trunk kept

upright, left and right arm lying alongside the body

and performing the movement, in a natural fashion.

In the anterior reaching and grasping the subjects

picked up, with the right hand, a cylinder (diameter,

3 cm; height, 6 cm; weight, 300 g) positioned on a

shelf in line (on the anterior direction) and at the

same height of the right shoulder, and returned the

cylinder to the starting position. The oblique and the

oblique opposite reaching and grasping tasks were

performed in the same manner of the anterior

reaching and grasping but with the object positioned

at ±45° with respect to the anterior direction. The

anterior trunk flexion was performed through a

maximal anterior flexion of the trunk. In the sit to

stand movement the subjects were seated

comfortably on a chair and got stand up in a natural

manner and at their preferred velocity. Ten cycles

were recorded for task and each person. Angular

excursion data were normalized to the movement

duration and reduced to 100 samples. For the

comparison of the Range of Motions (ROMs) we

used the root mean squared error (RMS):

()

oi di

RMS

−

∑

(1)

whereas the Coefficient of Multiple Correlation

(CMC), i.e. the positive square root of the adjusted

coefficient of multiple determination (Kabada et al.

1989, Steinwender et al. 2000) by means of the

following formula:

GLOBAL BIOMECHANICAL EVALUATION DURING WORK AND DAILY-LIFE ACTIVITIES

109

Figure 1: First column: pelvic angular behaviour in the sagittal plane (pelvic tilt); second column: pelvic angular behaviour

in the frontal plane (pelvic obliquity); third column: pelvic angular behaviour in the horizontal plane (pelvic intra-extra

rotation). 1

, 2

, 3

, 4

and 5

rows show the pelvic mean angular curves during the standing anterior reaching and

grasping, the standing oblique reaching and grasping, the oblique opposite reaching and grasping, the anterior trunk flexion

and the sit to stand respectively. In black and grey, curves acquired and computed by the high-quality optical motion

analysis system and the Wi-Fi transmission miniaturized device respectively. On x-axis and y-axis are reported percentage

cycle time duration and degrees respectively.

Table 1: ROMs (mean±standard deviation) calculated by the optoelectronic motion analysis system (Opt), by the wearable

inertial device (G+A), the paired t test P value, the RMS and the CMC for each motor task and for sagittal, frontal and

horizontal plane.

 Sagittal Frontal Horizontal

 Opt G+A P RMS CMC Opt G+A P RMS CMC Opt G+A P RMS CMC

AnteriorR&G 27.7±5.7 31.7±8.1 0.5 4.0 0.96 0.3±0.6 1.7±0.3 0.023 1.4 0.50 8.2±1.7 6.6±0.8 0.2 2.0 0.61

ObliqueR&G 23.5±1.5 26.9±1.8 0.066 3.4 0.94 6.3±4.6 3.3±1.2 0.336 5.2 0.82 9.9±2.0 9.7±1.5 0.897 1.6 0.96

ObliqueOppositeR&G 36.2±3.2 36.0±3.1 0.942 5.3 0.98 10.7±1.0 7.7±1.2 0.029 3.5 0.84 31.5±0.7 28.0±0.9 0.006 3.5 0.96

AnteriorTrunkFlexion 59.9±0.6 60.4±2.1 0.712 1.6 0.88 6.3±0.2 5.3±1.7 0.369 1.8 0.77 1.9±0.3 3.4±0.4 0.007 1.5 0.31

SeatToStand 30.6±2.0 31.6±1.6 0.536 3.4 0.99 1.9±0.4 4.0±1.0 0.013 2.1 0.70 ‐1.2±0.9 3.0±0.2 0.001 4.1 0.76

BIODEVICES 2010 - International Conference on Biomedical Electronics and Devices

110

()

∑∑

−

−=

tit

)1(

(2)

where T=100 (number of time points within the

cycle), N=2 (number of curves),

is the value at

the tth time point in the ith cycle,

y is the average

at time point t over N cycles:

∑

tit

(3)

and

is the grand mean of all

y :

∑∑

(4)

CMC was used to evaluate overall waveform

similarity of instantaneous angle curves: the closer

to 1 the CMC, the more similar the waveforms. The

statistical analysis was performed using SAS 8.2

(SAS Institute Inc., Cary, NC, USA). A paired t test

was applied in order to compare ROMs calculated

by the two techniques. P-values less than 0.01 were

considered statistically significant.

3 RESULTS

Results are summarized in Figure 1 and in Table 1.

, 2

, 3

, 4

and 5

rows of Figure 1 show the

pelvic mean angular curves during the standing

anterior reaching and grasping, the standing oblique

reaching and grasping, the oblique opposite reaching

and grasping, the anterior trunk flexion and the sit to

stand respectively. The 1

, 2

and 3

column of

Figure 1 show the pelvic angular behaviour in the

sagittal plane (pelvic tilt), in the frontal plane (pelvic

obliquity) and in the horizontal plane (pelvic intra-

extra rotation). In black and grey, curves acquired

and computed by the high-quality optical motion

analysis system and the Wi-Fi transmission

miniaturized device respectively. Table 1 shows the

ROMs (mean±standard deviation) calculated by the

optoelectronic motion analysis system (Opt), by the

wearable inertial device (G+A), the paired t test P

value, the RMS and the CMC for each motor task

and for sagittal, frontal and horizontal plane.

The results showed high similarity between the

extracted angle curves (Figure 1, Table 1) with

respect to the pelvic tilt on the sagittal plane

(CMC>0.88), the pelvic obliquity on the frontal

plane (CMC>0.70 except for the anterior reaching

and grasping) and the pelvic intra-extra rotation on

the horizontal plane (CMC>0.61 except for the

anterior trunk flexion). We also found low root mean

square errors in the computation of the

corresponding ROMs in the sagittal plane

(RMS≤5.3). Statistically significant differences in

the calculated ROMs were found only for the

standing oblique opposite reaching and grasping,

anterior trunk flexion and sit to stand on the

horizontal plane (P>0.01).

4 CONCLUSIONS

We compared data acquired and computed by two

different complementary technologies: a wearable

inertial device and an optoelectronic system. The

former allows a simple setup and outdoor

acquisitions (e.g. work environment); the latter

represent the kinematic gold standard acquisition

system but it is not simple to set up in work

environment. The use of wearable inertial devices

can be considered very useful when a simple

biomechanical human global approach is needed

(e.g. study of the human mechanical energy

expenditure, of the whole-body stiffness and of the

centre of mass behaviour). Our results suggest also

the use of these devices in work environment

applications, where specific segmental analyses are

needed, such as the study of the pelvic behaviour. In

these conditions, they yield good precision and

accuracy values on those measures of angular

components that present high magnitude of ROMs,

such as sagittal and frontal component of the our

study. Planes on which ROMs have low amplitudes

don’t show good similarity of curves and present

high root mean square errors.

Furthermore, this study suggests applying these

accurate, inexpensive and easy to use methods to

measure the kinematics of the pelvis during common

work and daily-life activities.

These devices could as well be used by a

biofeedback approach, which is widely considered

as a valid tool in various rehabilitation contexts

(Nelson, 2007). Kinematic-based (Van Vliet, 2006)

biofeedback frameworks have been proposed for a

routine inclusion in rehabilitation protocols. On the

other hand, there aren’t evidences on the use of these

devices in the return back to work environment of

workers after injuries. As these techniques share the

advantage of being suitable for workers' self-

administration, they may also be suitable for use in

GLOBAL BIOMECHANICAL EVALUATION DURING WORK AND DAILY-LIFE ACTIVITIES

111

telerehabilitation, which is not yet widespread

mainly due to the unavailability of specific devices,

validated protocols and appropriate operators'

educational programs. Indeed, the use of these

methods at home or at work may allow workers after

injuries to gain control over their own motor

recovery, to increase frequency and duration of

physical training and to improve personal

involvement and satisfaction in the rehabilitation

program. A miniaturized, wearable device for

kinematic biofeedback, interfaced with a

telerehabilitation platform, may improve the quality

of rehabilitation due to a faster getting back of

workers to their usual environments, with a

beneficial effect on quality of life, a minimization of

lost opportunity costs for employers, who can be

treated onsite reducing absence at work, and a

decrease of the economic burden for the healthcare

system. These techniques will be relevant for the

National Health Service in order to provide data

about the feasibility of rehabilitation treatment

transfer from hospitals to work settings. Such

transfer may allow the National Health Service to

reduce costs and ameliorate the managing of the

resources employed in this context.

REFERENCES

Medved V. Measurement of human locomotion, Boca

Raton, USA: CRC Press, 2001.

Cappozzo A, Della Croce U, Leardini A et al., “Human

movement analysis using stereophotogrammetry. Part

1: Theoretical background”, Gait & Posture

2005;21:186-96.

Boonstra Miranda C., Rienk M.A. van der Slikke, Noe

L.W. Keijsers, Rob C. van Lummel, Maarten C. de

Waal Malefijt, Nico Verdonschot, “The accuracy of

measuring the kinematics of rising from a chair with

accelerometers and gyroscopes”, Journal of

Biomechanics 39 (2006) 354–358.

Coley Brian, Brigitte M. Jolles, Alain Farron, Aline

Bourgeois, Francois Nussbaumer, Claude Pichonnaz,

Kamiar Aminian, "Outcome evaluation in shoulder

surgery using 3D kinematics sensors”, Gait & Posture

25 (2007) 523–532.

Lau Hongyin, Kaiyu Tong, “The reliability of using

accelerometer and gyroscope for gait event

identification on persons with dropped foot”, Gait &

Posture 27 (2008) 248–257.

Plamondon A., Delisle A., Larue C., Brouillette D.,

McFadden D., Desjardins P., Lariviere C., “Evaluation

of a hybrid system for three-dimensional measurement

of trunk posture in motion”, Applied Ergonomics 38

(2007) 697–712.

Pfau Thilo, Thomas H. Witte and Alan M. Wilson, “A

method for deriving displacement data during cyclical

movement using an inertial sensor”, The Journal of

Experimental Biology 2005, 208, 2503-2514.

Veltink Luinge H.J, P.H., Batenc C.T.M., “Ambulatory

measurement of arm orientation”, Journal of

Biomechanics 40 (2007) 78–85

Zijlstra Agnes, Jon H.M. Goosen, Cees C.P.M. Verheyen,

Wiebren Zijlstra, “A body-fixed-sensor based analysis

of compensatory trunk movements during

unconstrained walking”, Gait & Posture 27 (2008)

164–167.

Ferrigno, G. and Pedotti A., “ELITE: a digital dedicated

hardware system for movement analysis via real-time

TV signal processing”, IEEE Trans. Biomed. Eng. 32

(1985) 943–950.

Kabada M.P., Ramakrishnan M.E., Gainey W.J., Gorton

G. and Cochran G.V.B., “Reproducibility of

kinematic, kinetic, and electromyographic data in

normal adult gait. Journal of Orthopedic Research, 7

(1989) 849-860.

Steinwender G., Saraph S., Scheiber S., Zwick E.B. and

Uitz C., “Intrasubject repeatability of gait analysis data

in normal and spastic children”, Clinical

Biomechanics, 15(2) (2000), 134–139.

Nelson LA, “The role of biofeedback in stroke

rehabilitation: past and future directions”, Topics in

Stroke Rehabilitation, 14 (2007) 59-66.

Van Vliet PM, Wulf G., “Extrinsic feedback for motor

learning after stroke: What is the evidence?”,

Disability and Rehabilitation, 28 (2006) 831-840.

BIODEVICES 2010 - International Conference on Biomedical Electronics and Devices

112