Wearable Optical Sensor For Low Back Pain

Monitoring
Erna B. Tamizi M. A. Zawawi *corresponding author Roszitaida A.
Faculty of Electrical & Electronics Faculty of Electrical & Electronics Faculty of Electrical & Electronics
Engineering Technology Technology Technology
University Malaysia Pahang University Malaysia Pahang University Malaysia Pahang
Pekan, Malaysia Pekan, Malaysia Pekan, Malaysia
Email : ernatamizi@gmail.com Email : mohdanwar@ump.edu.my Email : rose8474@gmail.com

Abstract—Low back pain (LBP) is one of the major issues
related to health problems among adults especially people of
working age range. Manual therapy is one of the approaches in
treating LBP, aside from exercise, biofeedback, relaxation,
massage, and rehabilitation. It is often conducted by
physiotherapist or chiropractor to encourage joint mobilization
of the spine under a controlled environment. To effectively
monitor the progress of spine recovery in LBP treatment,
certain device is required to assist the physiotherapist in
collecting related static and dynamic spine data. Conventional
ways to get the spine data are by using X-ray and MRI, but they
are less preferable for repetitive use due to expensive and may
cause skin irritation. Alternative devices for this application
include goniometer, inclinometer, tape measure as well as spine
mouse. Although these options are low cost, manual handling of
the devices by inexperienced users will result in inaccurate spine
data collection. The aim of this paper is to study and compare
the performance of several spine monitoring devices that have
been developed by past researchers and then to propose an
alternative solution of spine monitoring sensor based on optical
sensor that could provide continuous data of the spine
kinematics. A sensor design using a microbending optical fibre
technique could potentially give various advantages because of
the small sensor size and low cost needed for sensor component
assembly which will only use a visible wavelength LED and
photodiode, and less affected to external signal interference.
history and physical examinations are crucial in order to
classify the problem background. During the physical
examination, a physiotherapist will measure the range of
motion (ROM) of the lumbar spine. There are three main
points to measure the ROM and the normal range are shown
in Fig. 1.
During the assessment, a physiotherapist will use a
goniometer to measure the ROM of each lumbar in flexion
(forward), extension (backward), and the lateral (side)
bending. As shown in Fig. 2, goniometer is one of the
conventional instruments that is typically used to measure the
range of motion of human joints[3]. However, this tool has
some limitations in measuring the ROM as the reading
accuracy is highly depending on the examiner’s prior
experience and thus human error factor may occur[4].
Besides, the same goniometer should always be used for the
same target applications in order to reduces chances of
instrumental error[3].

Keywords—wearable system, optical fibre sensor, plastic

optical fibre, spine monitoring

I. INTRODUCTION
Lower back pain (LBP) is usually an excruciating
condition that is experienced both in general and clinical
treatment. According to Ministry of Health (MOH) Malaysia,
the prevalence of LBP in most countries are between 10-63% Fig.1. Physical assessment and the normal Range of Motion (ROM)
while in Malaysia the rate is between 12-60%[1]. Low back for low back[5].
pain is described as pain in the lumbosacral region, buttocks,
and/or thighs. It is a painful condition where it can affect our
quality of life and it can be harmful if left untreated. For some
people, LBP will heal spontaneously without any treatment.
This kind of LBP might be caused by a ligament or muscle
sprain. On the other hand, one can suffer an acute or severe
back pain which may lead to chronic back pain if no
systematic intervention is carried out. If the pains go on for
about less than 12 weeks before it heals, it is classified as an
acute back pain. Meanwhile, if the pain continues after 12
weeks, it is then classified as a chronic back pain and it must
be treated immediately[1].
The initial state of examining a LBP patients are by doing
a history and physical examination towards the patients. As Fig.2. Goniometer
stated in The Malaysian Low Back Pain Guideline[2], both
 A rather different approach using the same inertial sensor
was also demonstrated in 2003 by using a ceramic
gyroscope[9]. In this design, two sensors were place along the
lumbar spine; one at the L1 lumbar spine segment and the
second sensor over the sacrum. Both sensors were attached
using a plastic screw to a small, mouldable plastic plate. The
ceramic vibrating unit used in this paper is ultra-small,
lightweight, and easy to apply and have a quick response.
These features make them ideal for joint motion measurement.
However, the zero-frequency offset when the sensor is
stationary is the main limitation of the gyroscope. This offset
Fig.3. Measuring the lumbar flexion using a Goniometer will then produce signal drift at the output reading of the
sensor.
Fig. 3 shows how to measure the lumbar flexion using the
goniometer[6]. Firstly, patient must be at a straight position In another paper, Mattman et al.[10] used strain sensors
with both hands hugging the body. The goniometer will then that were attached to the back of a tight-fitting clothing. The
be placed by a physiotherapist at the lowest level of the rib and sensors will measure the strain produced by the tight-fitting
reading will be taken for the starting position. The ROM of cloth whenever the wearer moved. The type of sensor
lumbar flexion is measured at the end position with the embedded in the garment is a thread type sensor which can be
goniometer stays at the same point. The main downside of attached to the textile using a silicone film as illustrated in
goniometry is the starting position, and the long axis of the Fig. 5[10]. These sensor threads have an elasticity range of
limb can only be visually estimated[7]. This can affect the 100% so it is suitable to be used on the tight-fitting garment.
measurement gain from the same patient. For chronic low Even though sensor concept is interesting with lightweight
back pain patient, X-ray and Magnetic resonance imaging characteristic, the use of standard garment size could lead to
(MRI) are the common tests used in current clinical settings. limited applicability to wearer with different body sizes.
However, both tests are not suitable for repetitive used as it
Another method that are previously used in monitoring the
may cause skin irritation due to the high radiation level and it
lumbar spine movement is the motion-based sensor. In this
is also costly.
case, an electromagnetic monitoring system is used in
This paper will study and compare recently available types detecting motion as it allow an absolute and comparative
of devices and their respective sensing approach in monitoring translational and orientation data to be obtained in all three
the ROM of lumbar spine. Then, the advantages and dimensions[11]. This system can be used for a real-time
disadvantages of each sensor will be discussed and before a monitoring with no line-of-sight problems. Some populate
new sensor design based on optical fibre technology will be option using an electromagnetic tracking system that has been
proposed for lumbar spine monitoring application. developed so far is FASTRAK. This device used miniature
light weight motion sensors that are attached on the skin at the
back of the body.
II. COMPARISON BETWEEN DIFFERENT SENSORS FOR
SPINE MONITORING APPLICATIONS
Human spine system can be divided into three main areas,
namely cervical, thoracic, and lumber, where cervical has 8
spinal cord segments, 12 for thoracic and 5 for lumber. Figure
4 shows the illustration of human spine and the nerve system
along the spinal cord. Lower back pain is a type of pain that
exists around the lumbar region which consists of 5 vertebrae
labelled as L1 – L5[1].
Due to the significant impact of LBP to human’s daily life,
various types of sensors have been developed related to the
lumbar spine application. In this section, we will discuss
several types of sensors and summarize the advantages and
disadvantages of each device. The first sensor applied for the
lumbar spine monitoring is based on the use of inertial
sensor[8]. A Lycra® rowing suit was designed and integrated
with piezoresistive arrays sensor and a tri-axial accelerometer
at the back of the suit. The piezoresistive arrays are made up
from an Electrical Conductive Elastrometers (CE) strip which
permits an unimpeded monitoring of body posture and
gesture. This skin-mounted type sensor performance,
however, is significantly influenced by the non-linearity,
hysteresis and long transient time issues. This limitation
caused a change in its resistance as the raw sensor output is
insufficient to represent the mechanical deformation due to the
human joint movement. Fig.4. Human spine structure and functions[12]

Fig.5. Sensor thread attached to the textile using a silicone film[10].
Approximately about four motions sensors are used in a
complete device setup where the angle and distance between
each sensor are being measured to obtain the desired
result[13]. This device has a good reliability and accuracy
when used to measure the static human posture but contrarily
when used in dynamic measurement. This is due to the change
of distance between each sensor that effected the accuracy of
the system. When the range increases, the accuracies of the
system decreases[14]. Other limitations of the system is that
the sensors themselves are very expensive as compared to the
other alternative devices[11].
The other sensor type that is widely used for human health
application nowadays is fibre-optic based sensors. Fibre optic
sensors can be categorised into several techniques, such as
intensity-based, wavelength-based, and phase-based sensors.
Optical based sensor is a popular alternative for sensor
development nowadays because of the reliability, immunity to
electromagnetic interference and it is relatively inexpensive.
In a study established by Williams et al.[15], it was shown that
fibre-optics are capable of providing a real-time feedback of
the lumbar motion which is an advantage for health
monitoring applications. The device is also easy to use without
needing any special setup which make it possible to be use for
patient whether under a clinical supervision or at home. In
another study, an intensity modulation approach was used in
designing an optical-fibre bending sensor[16]. The designed
sensor can measure the spine in the range between 0֯ to 20֯
which meets the typical range of motion of human spine[17].
This optical sensor was placed inside a plastic housing to
ensure a consistent sensor position during assessment, and the
tilting angle between the input and output fibres were used to
provide the continuous range of motion data of the spine.
III. REVIEW ON WEARABLE DESIGN REQUIREMENTS
Designing a good and reliable wearable device is an
important aspect in designing an accurate sensor for human
monitoring applications. In this section, we will review on the
design requirements needed in designing a comfortable and
reliable device. Wearable technologies have gained popularity
as compared to laboratory systems in allowing ambulatory
system for human motion analysis. To design a good wearable
device, there are certain points needed to be taken into
consideration. In a study established by Frances-Morcillo et
al.[18], there are two main points in designing a wearable
device which is human factor and human-computer
interactions. Under this section, we are going to focus on the
human factor issue. According to Canina et al. in 2008, to
design a good wearable device, we must sure it is comfortable
to the wearer, and it has a non-obstructive shape. These are
important because it make the wearers feel that the applied
device does not interrupt their daily activities[19]. This is
particularly important because, most of the time the wearable
devices need to be attached for quite a long time on the human
skin/body, thus the material must be comfortable, and its
weight need to be lightweight. In another study performed by
Andreoni et al.[20], the design factors include the physical
elements of the device such as structure, the rigidity of the
device cover and its associated components, the aspect ratio,
and other design-related issues in relation to anthropometry
and gender, spots on the body, elasticity, and adherence to the
body adhesive section or of the garments.
Besides the physical features listed above, the other thing
that should be taken into consideration is the cognitive
ergonomics. Tomberg et al.[21] in his study has stated that the
tolerance for error is an important features for a measuring
devices. It means a design can minimise the risks and the
negative effects of unintentional or involuntary behaviour.
This can give an accurate and precise reading from the device.
Other than tolerance for error, the other concept needed to be
implied is the simplicity of use. This concept design can help
built a wearable device that is easy to use, regardless of the
user's background, knowledge, language skills, or current
concentration level[21].
Emotional ergonomics is another requirement that is
needed to be considered in designing a wearable design related
to the human factor. Motti et al. stated that a subtle method
can assure more privacy and discretion to users[22]. It is
important to make sure that the user can use the device without
disturbing other nearby people. Some people are concern with
the excessively attracting people’s attention while using a
wearable device. Device’s responsiveness is important in
providing a real time monitoring device. Ensuring high
responsiveness helps users to achieve a more efficient and
productive result. Table 1 shows the main wearable design
requirements proposed in the paper related to the human
factor. All these characteristics are necessary to be applied
when designing a wearable device so that it can be use
comfortably by human. The table is categorized into three
types of ergonomics along with their respective requirements
as well as the parameters being used to specify each category.
IV. SENSOR COMPARISON FOR CURRENT SPINE SENSOR
The type of sensors used to assess spine movement varied
from strain gauges-based sensors, textile piezo-resistive
sensor, electromagnetic sensor, and tri-axial accelerometers.
Several papers are studied, and the sensors details are listed
in Table 2. In this table, we compared the sensor in term of
the sensor type, size and weight, sensor placement, and the
spine movement obtained from the sensor.
From the table, we can see that the size and weight of the
systems used was reported in only few studies. Sensor
placement reported in the studies are mostly placed over one
or more lumbar (L1, L3, L4, L5), sacral (S1, S2) and thoracic
(T12) segments. Velcro strap, single and double-sided
adhesive tape, elastic band, and neoprene strap are the various
type of fixation techniques introduced in the studies. Some
sensors were also attached using a rigid support or fixed into
clothing. However, in term of accuracy and reliability, it is
difficult to make a comparison as they were calculated based
on measurement taken under different environment and with
different test procedures. Table 3 shows the measured
accuracy and reliability of system described in each paper.
V. SENSOR DESIGN CONCEPT FOR OPTICAL-BASED
SENSOR SPINE MONITORING APPLICATION
Based on the reviewed articles, there are still few
alternatives to design a continuous spine monitoring device
that is based on wearable technology. The aim is not only to
allow wearable device capability, but also to overcome the
limitations of existing spine monitoring including, over size
and weight issues that cause discomfort to wearer, poor
output accuracy, significant output drift as well as does not
support continuous/ in situ measurement. The use of plastic
optical fibre can offer various advantages to the spine
monitoring sensor technology due as it comprises of small
sensor module, low cost for sensor component assembly by
using visible wavelength LED and photodiode, and less
affected to external signal interference. The problem with
signal drift can also be eliminated via suitable intensity
referencing method for optical fibres. The sensor will be
placed along L1 to S1 segments to get the best output of the
lumbar spine movement. However, it is important to note
that, the present of significant temperature variation may
have some effect on the overall output of the optical based
sensor due to LED and photodetector sensitivity to
temperature effect. Nevertheless, the use of referencing
technique such as bridging circuit can compress this
unwanted effect to the optical sensor output.

TABLE I. WEARABLE DESIGN REQUIREMENT[18]

Design Classification Design Requirements Design Parameters

Physical Ergonomics Safety Harm Anxiety

Durability Resistance
Comfort Shape

Breathability
Hygiene
Temperature

Sizing
Obstructiveness
Cognitive Ergonomics Usability Simplicity Intuitiveness
Reliability Precision Effectiveness
Emotional Ergonomics Engagement Privacy Subtlety

Aesthetics Fashion form language

TABLE II. COMPARISON BETWEEN DIFFERENT TYPES OF SPINE MONITORING SENSORS

Articles Type Size (mm) and Sensor Placement Spine Outcome Reported
Weight (g)
Bartalesi et al., Textile piezoresistive sensor Relatively small Piezoresistive strip over the Length of lumbar arch
2010[8] (single strip of conductive and lightweight lumbar spinous process L5/S1
elastometer) and 2 tri-axial Accelerometers: T12, Sacrum.
accelerometers
Lee et al., 2003 [9] 3D inertial tracking system (The Size = 26.9× Over the Spine range of movement
IS-300 PRO, InterSense Inc., 34.0× 30.5, L1 spinous process and the
Burlington, MA 01803) Weight = 59.5 Second sensor over the sacrum.
Mattman et al., 2007 Thread like novel strain sensor Diameter: 0.3mm, 21 strain sensors stitched on a Spine segment movement
[10] Length: 2cm garment and placed at the back of variation (mm)
the body.
Faber et al., Full body inertia/magnetic motion Relatively large Pelvis, Head, upper arms, 3D L5/S1 moment time series
2020[23] capture system and heavy forearm, thighs, shank, feet, and peaks; ground reaction
scapulae, sternum, hand forces
Walgaard et al., Single inertia sensor comprised of Relatively large T1/2, T12, S1 (Elastic garment) Average angles of thoracic and
2016[24] 3 accelerometers and 3 and heavy lumbar curves in the sagittal
gyroscopes and coronal planes
Williams et al., A series of 8 paired fibre-optic Size = From L1 to S1 Lumbar curvature (static
2010[15] sensors, attached to a ribbon of 480×13×1.3 posture)
sprung steel (L×W×H)
M.A. Zawawi et al., Intensity-based optical fibre (fibre Size = 80×12×8 Between L1 to L4 (random Spine range of motion (angle)
[16] tilting angle) (L×W×H) positions)
 TABLE III. ACCURACY AND RELIABILITY OF SYSTEM
Referred Articles & Sensor type Accuracy/Reliability

Bartalesi et al., 2010[8] 2% error in length estimation;

(piezoresistive) High correlation greater than 0.8 when comparing lumbar arch from the newly developed wearable
system with the stereophotogrammetric system.
Lee et al., 2003 [9] Mean error was found to be 0.81 ± 0.14°, and the system was considered sufficiently accurate for the
(inertial) measurement of movements of the lumbar spine.
Mattman et al., 2007 [10] Measurement error was ±3.5% over a strain range of up to 100%.
(strain sensor)
Faber et al., 2020[23] RMS errors were below 10 Nm for flexion, lateral flexion and twist L5/S1 moments time series. R^2
(inertial/magnetic) values > 0.993 for flexion L5/S1 moment and below 0.993 for lateral flexion and twist L5/S1 moments.
ICC of the absolute peak moments were 0.971, 0.781 and 0.69 for flexion, lateral flexion and twist
respectively. 3D GRF RMS errors remained below 20 N, R^2 above 0.981 for vertical GRF, around
0.6 for anterior/posterior and medio/lateral GRF. ICC of the vertical GRF peaks between systems was
0.998, and 0.948 and 0.559 for anterior/posterior and medio/lateral GRF peaks. During fast trunk
bending overestimation of about 15 N in vertical GRF peak.
Walgaard et al., 2016[24] RMSE < 10% for 3D lumbar accelerations, velocities, displacements, and angles between systems
(accelerometer & gyro) except for sideway displacement, and non-sagittal plane rotation with RMSE 40.1 ± 47.4%. ICC ≥
0.867 for lumbar flexion range of motion, maximum flexion velocity, maximum forward velocity,
forward velocity during seat-off and heel strike, maximum vertical velocity, except for vertical velocity
at heel strike ICC = 0.649. Mean absolute differences were 0.45 ± 0.35֯ for flexion range, 16.9 ± 16.6֯
for maximum flexion velocity and 0.1 ± 0.06 m/s or lower for other velocity measures.
Williams et al., 2010[15] The accuracy of the fibre-optic system as measured by the absolute mean difference for the peak
curvature measured by the two systems for the whole lumbar spine was 2.3±2.3◦ and 2.5±2.7◦ and for
the lower lumbar spine was 2.7±2.2◦ and 2.3±2.0◦ for flexion and lifting.
M. A. Zawawi et. al,[16] Resolution = < 2◦ in flexion/extension direction)
Drift = <0.25% for 2 hours measurement under consistent temperature
Accuracy was not specified at the time of publication.

For the proposed sensor design, we can conclude that the
best material in making a wearable device for the spine
should be adjustable yet sturdy at the sensor placement. A
back strap belt design is chosen as the core. This design is
selected to meet the physical ergonomics requirement as
stated in Table 1. The strap can be discreet since it can be
worn under clothes and can be worn by anyone with different
sizes. It is light and comfortable and allows us to move round
without limitation. An elastic band will be use as the belt of
the proposed design. The sensor will be attached on a
bendable aluminium strip as illustrated in Fig. 6 and
embedded to the elastic band right at the back of the body
along the spine. Bendable aluminium strips are soft and
flexible. Aluminium has such a good elongation that the strips
can be bent to adapt to varied installation conditions. It can
hold the sensor solidly and at the same time can follow the
movement of lumbar spine. With this proposed design, all
three wearable design requirements indicated in Table 1 are
fulfilled. Fig. 6 shows the proposed wearable device design
for the spine monitoring system.
For the sensor modulation method proposed for the sensor
design, the use of microbending technique could potentially
give an advantage of small dimension and lightweight sensor,
as long as the ‘teeth-like’ material and its shape that will be
used to influence the optical fibre curvature radius and
carefully design to optimize the sensitivity of the sensor with
respect to spine motion without compromising the wearer
comfort.
Fig. 7 shows the design concept of the proposed sensor
using a microbending optical fibre technique. Microbending
concept is the bending of an optical fibre in micro-size and
modifies the optical power loss. This bending is caused by
the pressure on the optical fibre in the axial direction. If the
optical fibre is bent in axial direction of the optical axis, the
transmission coefficient of the light will change with a
magnitude as the following equation[25]:
∆T
∆T = A𝑝 kf −1 ∆P ()
∆X

where ∆T/∆X are coefficient which relates to the change in

transmission of light in a deformed fibre, A𝑝 is the deformer
area, and ∆𝑃 are the changes in pressure applied to the sensor.
Elastic
band Aluminium strip Table 4 shows the bending angles of the lumbar spine, for
lateral flexion, flexion, and extension movements that ranges
between 1֯ to 13֯ and it is expected to be measured up to 20֯
with the use of the microbending optical fibre sensor.

Fig.6. The proposed design of wearable setup for spine monitoring

 TABLE IV. RANGES OF SEGMENTAL MOTION FOR DIFFERENT ACKNOWLEDGEMENT
MOTION TYPES[17]
The authors would like to thank the Ministry of Higher
Level of Lumbar Mean Range (o) Education for providing financial support under Fundamental
Spine Research Grant Scheme (FRGS) No.
Lateral Flexion Flexion Extension
FRGS/1/2019/TK04/UMP/02/18 (University reference
L1-2 5 8 5
RDU1901218) and Universiti Malaysia Pahang for
L2-3 5 10 3 laboratory facilities as well as additional financial support
L3-4 5 12 1 under Internal Research grant RDU1901218.
L4-5 3 13 2
L5-S1 2 9 5
REFERENCES

[1] R. 45 and September, “Normal Spinal Anatomy,”

[2]
[3]
[4]
Teeth-like material for spine application
[5]
[6]
[7]
Fig.7. The proposed design of microbend sensor for spine monitoring
VI. CONCLUSION
Spine monitoring device can be very useful in assessing a
LBP patient. This can ensure an accurate measurement for
each patient no matter who in charge of the patient. A
continuous monitoring also can be taken for a chronic LBP [8]
patient during rehabilitation. Several types of spine
monitoring devices have been reviewed and compared to get
the best possible design to be implemented in the future.
Some of the device need to be done some modification in
term of the design and the type of sensor used.
[9]
The potential use of a microbending plastic optical fibre
in developing a spine monitoring device can offer various
advantages because of the small sensor size and low cost
needed for sensor component assembly which will only use a
visible wavelength LED and photodiode, and less affected to [10]
external signal interference. However, choosing the right type
and size of the fibre optics is crucially important in order get
a good measurement result from the developed design.
[11]
vol. 79, no. 1. 2007.
A. Malik, M. Hussein, M. Mal, and F. Glas, “The
Malaysian Low Back Pain Guideline,” FEBS J., vol.
281, no. 9, p. 2313, 2014, [Online]. Available:
http://www.ncbi.nlm.nih.gov/pubmed/24786108.
“Goniometer - Physiopedia.” .
R. M. F. de Carvalho, N. Mazzer, and C. H. Barbieri,
“Analysis of the reliability and reproducibility of
goniometry compared to hand photogrammetry,”
Acta Ortopedica Brasileira, vol. 20, no. 3. pp. 139–
149, 2012, doi: 10.1590/S1413-
78522012000300003.
J. Medlej, “Human Anatomy Fundamentals:
Flexibility and Joint Limitations,” Envatotuts. 2014,
[Online]. Available:
https://design.tutsplus.com/articles/human-anatomy-
fundamentals-flexibility-and-joint-limitations--
vector-25401.
“MEASUREMENT of RANGE of MOTION of the
THORACIC and LUMBAR SPINE |
Musculoskeletal Key.” [Online]. Available:
https://musculoskeletalkey.com/measurement-of-
range-of-motion-of-the-thoracic-and-lumbar-spine/.
S. Nussbaumer, M. Leunig, J. F. Glatthorn, S.
Stauffacher, H. Gerber, and N. A. Maffiuletti,
“Validity and test-retest reliability of manual
goniometers for measuring passive hip range of
motion in femoroacetabular impingement patients.,”
BMC Musculoskeletal Disorders, vol. 11. 2010, doi:
10.1186/1471-2474-11-194.
R. Bartalesi, F. Lorussi, D. De Rossi, M. Tesconi, and
A. Tognetti, “Wearable monitoring of lumbar spine
curvature by inertial and e-textile sensory fusion,”
2010 Annu. Int. Conf. IEEE Eng. Med. Biol. Soc.
EMBC’10, pp. 6373–6376, 2010, doi:
10.1109/IEMBS.2010.5627294.
R. Y. W. Lee, J. Laprade, and E. H. K. Fung, “A real-
time gyroscopic system for three-dimensional
measurement of lumbar spine motion,” Med. Eng.
Phys., vol. 25, no. 10, pp. 817–824, 2003, doi:
10.1016/S1350-4533(03)00115-2.
C. Mattmann, O. Amft, H. Harms, G. Tröster, and F.
Clemens, “Recognizing upper body postures using
textile strain sensors,” Proc. - Int. Symp. Wearable
Comput. ISWC, pp. 29–36, 2007, doi:
10.1109/ISWC.2007.4373773.
S. A. Sobey, “Three-Dimensional Measurement of
 Spinal Kinematics and Whole-Body Activity [24] S. Walgaard, G. S. Faber, R. C. van Lummel, J. H.
Recognition,” no. July, 2019. van Dieën, and I. Kingma, “The validity of assessing
[12] X. Y. Gu, S. Ren, Y. Shi, N. Wang, Z. H. Tong, and temporal events, sub-phases and trunk kinematics of
M. L. Cai, “Restoration Methods of Respiratory the sit-to-walk movement in older adults using a
Function for Spinal Cord Injury,” Math. Probl. Eng., single inertial sensor,” J. Biomech., vol. 49, no. 9, pp.
vol. 2020, 2020, doi: 10.1155/2020/7398789. 1933–1937, 2016, doi:
[13] K. Jordan, K. Dziedzic, P. W. Jones, B. N. Ong, and 10.1016/j.jbiomech.2016.03.010.
P. T. Dawes, “The reliability of the three-dimensional [25] S. N. Kane, A. Mishra, and A. K. Dutta, “Preface:
FASTRAK measurement system in measuring International Conference on Recent Trends in
cervical spine and shoulder range of motion in Physics (ICRTP 2016),” J. Phys. Conf. Ser., vol. 755,
healthy subjects,” Rheumatology, vol. 39, no. 4, pp. no. 1, 2016, doi: 10.1088/1742-6596/755/1/011001.
382–388, 2000, doi:
10.1093/rheumatology/39.4.382.
[14] Polhemus, “Fastrack 3d digitizer & quad sensor
motion tracker,” pp. 1–2, 2017, [Online]. Available:
https://polhemus.com/_assets/img/FASTRAK_Broc
hure_1.pdf.
[15] J. M. Williams, I. Haq, and R. Y. Lee, “Dynamic
measurement of lumbar curvature using fibre-optic
sensors,” Med. Eng. Phys., vol. 32, no. 9, pp. 1043–
1049, 2010, doi: 10.1016/j.medengphy.2010.07.005.
[16] M. A. Zawawi, S. O’Keeffe, and E. Lewis, “Plastic
optical fibre sensor for spine bending monitoring
with power fluctuation compensation,” Sensors
(Switzerland), vol. 13, no. 11, pp. 14466–14483,
2013, doi: 10.3390/s131114466.
[17] N. Bogduk, “Clinical Anatomy of the Lumbar Spine
and Sacrum,” no. 3, p. 2000, 2000.
[18] L. Francés-Morcillo, P. Morer-Camo, M. I.
Rodríguez-Ferradas, and A. Cazón-Martín,
“Wearable design requirements identification and
evaluation,” Sensors (Switzerland), vol. 20, no. 9,
2020, doi: 10.3390/s20092599.
[19] M. Canina and V. Ferraro, “The biodesign approach
to wearable devices,” Proc. 5th Int. Work. Wearable
Implant. Body Sens. Networks, BSN2008,
conjunction with 5th Int. Summer Sch. Symp. Med.
Devices Biosensors, ISSS-MDBS 2008, pp. 264–267,
2008, doi: 10.1109/ISSMDBS.2008.4575070.
[20] G. Andreoni, C. E. Standoli, and P. P. Paolo,
“Defining requirements and related methods for
designing sensorized garments,” Sensors
(Switzerland), vol. 16, no. 6, 2016, doi:
10.3390/s16060769.
[21] V. Tomberg, T. Schulz, and S. Kelle, “Applying
universal design principles to themes for wearables,”
Lecture Notes in Computer Science (including
subseries Lecture Notes in Artificial Intelligence and
Lecture Notes in Bioinformatics), vol. 9176. pp. 550–
560, 2015, doi: 10.1007/978-3-319-20681-3_52.
[22] V. G. Motti and K. Caine, “Human factors
considerations in the design of wearable devices,”
Proc. Hum. Factors Ergon. Soc., vol. 2014-Janua,
no. December, pp. 1820–1824, 2014, doi:
10.1177/1541931214581381.
[23] G. S. Faber, I. Kingma, C. C. Chang, J. T. Dennerlein,
and J. H. van Dieën, “Validation of a wearable
system for 3D ambulatory L5/S1 moment assessment
during manual lifting using instrumented shoes and
an inertial sensor suit,” J. Biomech., vol. 102, p.
109671, 2020, doi:
10.1016/j.jbiomech.2020.109671.