1 Collaborative Forum 3: Suit Sizing for Optimal Fit EVA Technology Workshop 2020 February 20 th , 2020 Moderator: Elizabeth Benson (KBR) Panelists: Richard Rhodes (NASA), Han Kim (Leidos), Dr. Rachel Vitali (University of Michigan) National Aeronautics and Space Administration
21
Embed
Collaborative Forum 3: Suit Sizing for Optimal Fit
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
PowerPoint PresentationEVA Technology Workshop 2020
February 20th, 2020 Moderator: Elizabeth Benson (KBR) Panelists:
Richard Rhodes (NASA), Han Kim (Leidos), Dr. Rachel Vitali
(University of Michigan)
National Aeronautics and Space Administration
SESSION AGENDA
• Session Objectives (5 min) • Introduction of Panel (20 min)
– Richard Rhodes (NASA JSC) – Han Kim (NASA JSC) – Dr. Rachel
Vitali (University of Michigan)
• Questions and Discussion (1 hr)
2
• Encourage open communication between NASA and the greater EVA
community, regarding the complex topic of space suit sizing and fit
assessment
– Recent advances in suit sizing and fit assessment tools – Current
challenges in suit sizing and fit assessment – The potential for
unique sizing and fit challenges on the lunar surface
3
Virtual fit checking of hardware What sizing and fit challenges are
posed by lunar surface operations in the suit?
Panelist – Richard Rhodes
EVA Technology Workshop 2020
February 20, 2020 Dr. Rachel V. Vitali TRISH Postdoctoral Fellow
University of Michigan, Ann Arbor
National Aeronautics and Space Administration
Examples of Previous Work using IMUs for Suit Assessment
• Di Capua and Akin (2012) – Comparing joint kinematics for three
different suits using joint angles
provided by IMUs and optical motion capture • Bertrand, Anderson,
Hilbert, and Newman (2014)
– Compared IMU-derived shoulder and elbow ranges of motion of the
human to those of the suit (both unpressurized and pressurized) to
study human-suit interactions
• Fineman, McGrath, Kelty-Stephen, Abercromby, and Stirling (2018)
– Quantified task specific knee range of motion and relative
movement
between human and suit for participants wearing an EVA suit with
different indexing conditions
Designing an Inertial Motion Capture Study
• Depending on specificity of the biomechanics you’re interested in
studying dictates how many sensors you will need to use
– An IMU at the sacrum frequently treated as a proxy for center of
mass dynamics
– IMUs attached to feet provide valuable insight into gait
parameters – IMUs attached to either side of a joint can provide
joint kinematics
• Depending on the task, experimental protocol usually dictates
what is possible when processing the data
Additional Considerations for Inertial Motion Capture
• Sensor calibrations (usually handled by manufacturer) •
Integration drift error: accelerometer and gyro errors integrated
into linearly (quadratically)
increasing errors in velocity (position) and orientation,
respectively • IMU orientation estimation: fusing independent
estimates of orientation from gyro, accelerometer,
and usually magnetometer to relate IMU frame to a world frame –
Many commercial products offer proprietary software that will
provide estimates – Ongoing efforts to improve methods
• Establishing a common world frame can be difficult – Magnetic
interference usually manifests differently
• IMU-to-body-segment calibration – No convention for defining
anatomical frames for inertial motion capture
• Synchronization sampling across an IMU array – Some commercial
products offer synchronized data logging
9
EVA Technology Workshop 2020
National Aeronautics and Space Administration
10
Medium Size HUT
Large Size
Extra-Large Size
Linear Measurement Based 3-D Scan & CAD Fit Check Boundary
Manikin Tests
• Shuttle EMU: Linear measurements were compared between the body
and suit • Z-2: A limited number of 3-D body scans were overlaid to
check the overlap the suit CAD • Z-2 & Z-2.5: Increased number
of body scans to assess “worst-case” fit testing (”boundary
manikins)
Tall
Short
h1
Suit-to-Body Overlap Score Calculation Machine Learning
Classifier
• Overlay the 3-D body scans with the CAD model of the suit •
Estimate the suit-to-body contact and overlap • Build a statistical
classifier to predict the fit probability as a function of the
suit-to-body overlap
3-D Body Scan
Fit Unfit
Suit Geometry
Lower Score High ScoreOverlap Score
Subject A Subject B
y
1.0
0.0
0.5
• Sort the potential subjects by overlap score and visually inspect
the overlap charts • From overlap charts, subjects “obviously
likely” to fit (or unfit) were excluded from physical testing •
Physical fit tests performed with borderline fit subjects •
Iteratively update the fit classifier by physical fit test
outcome
Subjects with Smaller Overlap Excluded from physical testing
Subjects with Larger Overlap Excluded from physical testing
Borderline Fit Subjects Selected for physical testing
Classifier to Estimate Crew Population Accommodation
13
• Project the classifier model to a large population database (US
Army; 3,890 Males, 1,712 Females) • Count fit vs. unfit cases and
estimate the accommodated proportion of the crew population • This
method enables identifying marginally fitting cases, i.e.,
Prob(Fit) ≈ 0.5 and fit surface gradient • This new information can
help to identify the design issues and iteratively optimize the
suit design
US Army Scan Database
Pre-Screened Cases (“Obviously Fit”)
Pre-Screened Cases (“Obviously Unfit”)
e
• Suit-to-body overlap is a key metric, but the specific magnitude
of acceptable overlap is still unknown • This study directly
measured the maximum tolerable depth of overlap by maximally
“pushing” a probe • Developed a parametric model and the outcome
was compared to the virtual fit tests
Subjective Reporting of Suit-to-Body Contact
15
Compare Outcome to the Virtual Suit-to-Body Overlap
• Physically tested subjects also reported the perceived locations
of suit-to-body contacts • The subjective reporting was compared to
physical and virtual suit contact and overlap
Discussion Topics
Topic 1 – Lunar EVAs
• What sizing and fit challenges are likely to be unique to
planetary suits working in partial gravity?
• How will forceful exertions and extreme postures influence
fit?
17
Subject crawling in prototype planetary suit (1-g)Subject walking
on reduced gravity aircraft
Topic 2 - Suit-Body Contact Assessment
- How do we discriminate ‘good’ vs. ‘bad’ suit-body contact? - How
do the anatomical properties of the contact location change suit
fit
(for example, bone vs. soft tissue)
18
Topic 3 – Custom vs. modular sizing
• Why do we have modular suits, and not custom-fitted suits that
are unique to each crewmember?
19
Apollo era: Custom sewn suits Shuttle Era: Modular suit
architecture
Topic 4 – Other disciplines
• What are examples of other fields that have similar fit and
sizing challenges, and how have they worked to resolve these
issues?
20
Designing an Inertial Motion Capture Study
Additional Considerations for Inertial Motion Capture
Slide Number 9
New Method: Large-Scale Testing for Virtual Suit Fit
Test Subject Selection and Iterative Classifier Training
Classifier to Estimate Crew Population Accommodation
Skin Compression Tolerance
Discussion Topics
Topic 3 – Custom vs. modular sizing
Topic 4 – Other disciplines