Skip to main content

Table 3 List of the selected 22 articles organised by first author and containing information related to the validation of wearable sensors for the measurement of joint angle in upper limb when compared to a three-dimensional motion analysis system

From: Measurement of Upper Limb Range of Motion Using Wearable Sensors: A Systematic Review

First Author

Aim of the study

Brand of Sensors

Sensor fusion algorithm

Placement of sensors

Comparison system

Used same segment tracking

Task(s)

Anatomical Segment(s)

Degrees of Freedom

Movements

Mean error (SD)

RMSE

Correlation coefficients

Calibration

System

Static

Dynamic

Robert Lachaine et al. [24]

Validate protocol

Xsens

KF

S1: Upper arm

S2: Forearm

S3: Hand

Optotrak

Yes

Elbow flex/ext, pro/sup; wrist flex/ext, ul/rad deviation, rotation and manual handling tasks

Shoulder

Elbow

Wrist

3DOF

3DOF

3DOF

Flex/ext

Ab/ad

Rotation

Flex/ext

Ab/ad

Pro/sup

Flex/ext

Rad/ul dev

Rotation

Optotrak ISB to Xsens ISB

–

✓

–

–

3.0°

2.9°

2.5°

2.9°

2.0°

2.6°

3.8°

2.8°

3.6°

–

Ligorio et al. [69]

Validate calibration method

YEI technology

–

–

Vicon

Yes

Flex/ext and pro/sup

Elbow

2DOF

2DOF 2DOF

Flex/ext

Pro/supFlex/ext

Pro/supFlex/ext

Pro/sup

Method A

–

✓

✓

–

8.5–11.1°

11.9–13.3°

–

Method B

–

3.4–3.6°

6.8–7.6°

–

Method C – Proposed

–

3.1–3.3°

3.8–4.0°

–

Fantozzi et al. [45]

Validate protocol

Opal

KBF

S1: Flat portion of the sternum.

S2: Laterally on the humerus above the centre and posteriorly.

S3: Distal forearm above the ulnar and radial styloid.

S4: Back of the hand.

Stereo-photogrammetric system (SMART-DX 7000)

Yes

Simulated front crawl

Shoulder

Elbow

Wrist

3DOF

2DOF

2DOF

Flex/ext

Ab/ad

In/ext rot

Flex/ext

Pro/sup

Flex/ext

Rad/ul dev

–

5.0° (4–6)

10.0° (7–11)

7.0° (5–8)

15° (12–17)

10.0° (7–11)

5.0° (4–5)

3.0° (2–4)

0.99

0.97

0.99

0.95

0.93

0.95

0.90

–

✓

–

Simulated breaststroke

Shoulder

Elbow

Wrist

3DOF

2DOF

2DOF

Flex/ext

Ab/ad

In/ext rot

Flex/ext

Pro/sup

Flex/ext

Rad/ul dev

–

5.0° (3–7)

3.0° (3–4)

8.0° (5–10)

6.0° (5–10)

5.0° (4–7)

4.0° (3–5)

-

0.99

0.99

0.98

0.97

0.98

0.93

Gil-Agudo et al. [30]

Validate protocol

Xsens

KF

S1: Trunk

S2: Back of the head

S3: Right arm

S4: Distal forearm

S5: Hand.

CODA

Yes

Shoulder rot, flex/ext and ab/ad; elbow flex/ext and pro/sup, wrist flex/ext and ul/rad deviation.

Shoulder

Elbow

Wrist

3DOF

2DOF

2DOF

Flex/ext

Ab/ad

In/ext rot

Flex/ext

Pro/sup

Flex/ext

Rad/ul dev

0.76° (4.4)

0.69° (10.47)

0.65° (5.67)

0.54° (2.63)

5.16° (4.5)

3.47° (9.43)

2.19° (4.64)

–

–

–

–

✓

Miezal et al. [15]

Validate sensor fusion/algorithm

Xsens

EKF, WLS

Not described

Natural Point Optitrack system 13 cameras

Yes

Eight-shaped movements at varied speeds, smooth parts imitating reaching and steering in the case of real-slow, and agile parts with quick starts and stops, as well as, parts reminding of sportive movements, such as boxing, in the case of real fast

Shoulder

Elbow

Wrist

1DOF

1DOF

1DOF

–

Chaintracker (real fast w/mag)

–

✓

–

9.38° (5.79)

11.91° (6.27)

7.37° (4.60)

–

–

Shoulder

Elbow

Wrist

1DOF

1DOF

1DOF

–

Chaintracker (real slow w/mag)

4.76° (2.24)

8.83° (4.64)

4.72° (2.61)

–

–

Shoulder

Elbow

Wrist

1DOF

1DOF

1DOF

–

Optitracker (real fast w/mag)

1.88° (0.91)

2.22° (1.38)

2.28° (1.15)

–

–

Shoulder Elbow

Wrist

1DOF

1DOF

1DOF

–

Optitracker (real fast w/mag)

1.27° (0.81)

2.16° (1.35)

2.32° (1.37)

–

–

Lambretcht et al. [62]

Validate sensor fusion/algorithm

Custom

DMP algorithm

S1: Sternum

S2: Upper arm

S3: Distal forearm

S4: Hand

Optotrak

Yes

Reaching movements

Shoulder

Elbow

Wrist

3DOF

2DOF

2DOF

Azimuth

Elev

Int rot

Flex

Pro

Flex/Ext

Dev

–

4.9°

1.2°

2.9°

7.9°

1.5°

5.5°

2.6°

0.99

0.99

0.99

0.99

0.99

0.97

0.94

✓

–

–

Zhang et al. [34]

Validate sensor fusion/algorithm

Xsens

UKF

S1: Sternum

S2: Lateral side above the elbow S3: Lateral and flat side of the forearm near the wrist

BTS SMART-D optoelectronic tracking system

Yes

Move the upper limb arbitrarily.

Shoulder

Elbow

3DOF

2DOF

Flex/ext

Ab/ad

Int/ext rot

Flex/ext

Pro/sup

Independent Estimation

–

✓

–

0.070° (0.083)

0.023° (0.042)

0.061° (0.061)

0.052° (0.155)

0.321° (0.265)

0.11°

0.04°

0.08°

0.16°

0.41°

0.99

0.99

0.99

0.81

0.96

Shoulder

Elbow

3DOF

2DOF

Flex/ext

Ab/ad

Int/ext rot

Flex/ext

Pro/sup

Constraints method

0.040° (0.039)

0.013° (0.018)

0.029° (0.032)

0.046° (0.100)

0.155° (0.143)

0.05°

0.02°

0.04°

0.11°

0.21°

0.99

0.99

0.99

0.88

0.96

Shoulder

Elbow

3DOF

2DOF

Flex/ext

Ab/ad

Int/ext rot

Flex/ext

Pro/sup

Papers proposed method

0.028° (0.029)

0.007° (0.013)

0.035° (0.036)

0.054° (0.093)

0.168° (0.153)

0.04°

0.01°

0.05°

0.10°

0.22°

0.99

0.99

0.99

0.89

0.96

Morrow et al. [42]

Validate protocol

Opal

–

Bilateral:

S1: Lateral aspect upper arms

S2: Forearms

Raptor 12 Digital Real-time Motion Capture System

No

Peg transfer task using straight laparoscopic surgical instruments.

Shoulder

Elbow

1DOF

1DOF

Elevation

Flexion

3.0° (2.1)

2.2° (1.6)

6.8° (2.7)

8.2° (2.8)

–

–

✓

✓

Callegas-Cuerro et al. [59]

Validate protocol

Invensense MPU-9150

KF

S1: External arm aligned with the humerus.

S2: Between the radial styloid and ulnar styloid, aligned with external part of the hand.

Qualisys Oqus 5

No

Flex/ext

Elbow

1DOF

Flex/ext

< 3.0° to < 5.0°

2.44%

–

–

✓

✓

Meng et al. [56]

Validate protocol

Shimmer

KF

Not described

Vicon Mocap System

No

(1) Raise shoulder. (2) Move shoulder right then left. (3) Clockwise axial rotation to its max, then rotate the upper arm counter clockwise. (4) Elbow extension move into flexion.

Shoulder

Elbow

3DOF

2DOF

Flex/ext

Ab/ad

In/ext rot

Flex/ext

Pro/sup

0.50° (1.79)

0.18° (1.34)

0.16° (1.96)

1.86° (1.85)

1.22° (2.87)

1.85°

1.35°

1.96°

2.62°

3.12°

–

✓

–

–

Cifuentes et al. [81]

Validate protocol

Custom

–

S1: Arm

S2: Forearm

Optical tracking system

No

Reaching and grasping from the rest position with the forearm on the table, at angle of approximately 90° with respect to the arm before reaching and grasping an object, and then returning it to starting position.

Elbow

1DOF

Flex/ext

No discrete data reported only figures of continuous data

–

–

–

Muller et al. [22]

Validate sensor fusion/algorithm

Xsens

KF*

S1: Thorax.

S2: Lateral side of the arm

S3: Posterior side of the wrist

Vicon

No

(1) Flex/ext in a horizontal plane with the shoulder abducted 90° flex/ext in a sagittal plane while standing with the elbow close to the trunk. (2) Flex/ext in a sagittal plane with the spine bent forward 90° and the upper arm aligned horizontally and parallel to the ground sup/pro with the elbow flexed 90°

Elbow

Elbow

2DOF

2DOF

Flex/ext

Pro/sup

Flex/ext

Pro/sup

Proposed algorithm

✓

✓

–

–

2.7°

3.8°

–

Manual alignment

–

3.8°

8.7°

–

Bertomu-Motos et al. [55]

Validate sensor fusion/algorithm

Shimmer

EKF

S1: Shoulder

S2: Upper arm

Optitrack

No

The activity consisted of taking a box from the perimeter and placing it in the centre of the screen.

Shoulder

5DOF

Unclear

Without compensation Filter

–

–

–

5.24° (3.38)

0.5° (1.6)

3.6° (2.1)

1.8° (1.0)

1.60° (0.6)

–

–

Shoulder

5DOF

Unclear

Compensation filter

1.69° (2.1)

1.1° (0.8)

5.9° (2.3)

2.6° (1.7)

0.9° (1.2)

–

–

Karunarathne et al. [68]

Validate sensor fusion/algorithm

BioKin WMS

KF*

S1: Near the elbow

S2: Wrist

Vicon

No

Lifting a water bottle

Elbow

1DOF

Flex/ext

High-pass filte—gyroscope

–

–

–

–

10.18°

–

Elbow

1DOF

Flex/ext

Low-pass filter—accelerations

–

18.30°

–

Elbow

1DOF

Flex/ext

Tradition complementary filter

–

10.30°

–

Elbow

1DOF

Flex/ext

Adaptive complementary filter

–

8.77°

–

El-Gohary et al. [50]

Validate Sensor fusion/algorithm

Opal

UKF

S1: Upper arm

S2: Forearm

Vicon motion analysis system

No

Single movements: Shoulder flex/ext, ab/ad, Elbow flex/ext and forearm sup/pro.

Shoulder

Elbow

2DOF

2DOF

Flex/ext

Ab/ad

Flex/ext

Pro/sup

–

5.5°

4.4°

6.5°

0.95°

0.98

0.99

0.98

0.95

–

✓

 

Complex tasks: (1) touching nose and (2) reaching for door

Shoulder

Elbow

1DOF

1DOF

–

9.8°

8.8°

6.5°

5.5°

0.94

0.95

El-Gohary et al. [51]

Validate Sensor fusion/algorithm

Opal

UKF

S1: Between the shoulder and elbow

S2: Near the wrist

Eagle Analog System, Motion Analysis

No

Single movements at different speeds: Shoulder flex/ext, ab/ad, Elbow flex/ext, sup/pro

Shoulder

Elbow

2DOF

2DOF

Flex/ext

Ab/ad

Flex/ext

Pro/sup

Normal speed

–

–

–

–

–

0.97

0.94

0.92

0.96

Shoulder

Elbow

2DOF

2DOF

Flex/ext

Ab/ad

Flex/ext

Pro/sup

Fast speed

  

0.94

0.91

0.89

0.93

Perez et al. [39]

Validate sensor fusion/algorithm

Xsens

–

S1: Back

S2: 18 cm from acromion

S3: 25 cm from epicondyle

S4: 5.5 cm from distal radio-cubital joint.

BTS SMART-D optoelectronic tracking system

No

Single movements: Shoulder flex/ ext, horizontal ab/ad, and internal rotation. Elbow flex, pro/sup and wrist flex/ext.

Shoulder

Elbow

Wrist

3DOF

2DOF

1DOF

Flex/ext

Ab/ad

In rot

Flex

Pro/sup

Flex/ext

13.4°

17.2°

60.4°

5.8°

24.1°

11.6°

–

0.99

0.71

0.99

0.98

0.96

0.98

✓

–

–

Pouring water from a glass jar into a glass

Shoulder

Elbow

Wrist

3DOF

2DOF

1DOF

Flex/ext

Ab/ad

In rot

Flex/ext

Pro/sup

Flex/ext

13.8°

7.4°

28.8°

18.6°

11.7°

26.8°

–

0.99

0.90

0.85

0.97

0.92

0.92

Zhou et al. [37]

Validate sensor fusion/algorithm

Xsens

KF

S1: Lateral aspect of upper arm between the lateral epicondyle and the acromion process (5 cm from the AP)

S2: Wrist centre on the palmer aspect

CODA

No

Reaching, shrugging, forearm rotation

Elbow

2DOF

Flex/ext

Rot

0.4° (2.34)

0.06° (4.82)

2.4°

4.8°

–

–

✓

–

Luinge et al. [41]

Validate sensor fusion/algorithm

Xsens

KF

S1: Lateral upper arm near the elbow

S2: Dorsal side of the forearm near the wrist.

Vicon

No

(1) Mimicking eating routines (pouring a glass eating soup, eating spaghetti, eating meat, drinking). (2) Mimicking morning routines (splashing water on face and drying it using a towel, applying deodorant, buttoning a blouse, combing hair, brushing teeth).

Elbow

2DOF

–

No discreet data reported

–

✓

✓

Peppoloni et al. [57]

Validate kinematic model

Shimmer

UKF

S1: Scapula beside the angulus acromialis

S2: Lateral side of the upper arm above the elbow.

S3: Lateral side of forearm a few centimetres far from the wrist.

Vicon

No

Single movements:

Scapula elev/dep, ante-position/retro-position. Shoulder flex/ext, ab/ad, and int/ext rotation. Elbow flex/ext, pro/sup.

7DOF model

–

✓

✓

Scapula

Shoulder

Elbow

2DOF

3DOF

2DOF

Elev/dep

Prof/retr

Flex/ext

Ab/ad

In/ext rot

Flex/ext

Pro/sup

–

6.19°

3.43°

8.19°

10.68°

8.79°

5.00°

9.61°

0.65

0.74

0.94

0.63

0.97

0.99

0.85

5DOF model

Shoulder

Elbow

3DOF

2DOF

Flex/ext

Ab/ad

In/ext rot

Flex/ext

Pro/sup

–

7.03°

6.03°

4.95°

9.93°

11.29°

0.95

0.87

0.99

0.98

0.85

Robert-Lachaine et al. [25]

Validate calibration method

Xsens

KF

–

Optotrak

No

Single plane movements

–

–

–

No discrete data reported

–

–

–

Bouvier et al. [23]

Validate calibration method

Xsens

KF

S1: Sternum

S2: Central third of upper arm laterally (or slightly posterior)

S3: Dorso-distally on the forearm

S4: Dorsum hand

Eagle 4 Optoelectric system

No

Move through 9 calibration trials for each joint.

Shoulder

Elbow

Wrist

3DOF

2DOF

2DOF

Flex/ext

Ab/Ad

Wheel

Flex/ext

Pro/sup

Flex/ext

Ab/sd

–

–

–

–

–

–

–

–

–

20.46°

14.76°

14.21°

13.9°

–

–

–

0.84

0.94

0.93

0.68

–

✓

✓

  1. Abbreviations: RMSE root mean square error, SD standard deviation, CMC coefficient of multiple correlation, KBF Kalman-based filter, KF Kalman filter, EKF extended Kalman filter, UKF unscented Kalman filter, WLS weighted least squares, Flex flexion, Ext extension, Pro pronation, Sup supination, Ab abduction, Ad adduction, Dev deviation, Rad radial, Uln ulnar, In internal, Ex external, Rot rotation, Elev elevation, Dep depression, DOF degrees of freedom, C customised, M manufacture
  2. *The information was obtained from the manufacturer procedure manual or other referenced papers
  3. –Information was not reported and/or unclear in the study and/or unable to be obtained from the manufacturer manual