Towards Safer Human-Robot Interaction Assessing robotic safety using the Head Injury Criterion (HIC) and the Chamois Laceration Scale (CLS)

1
Towards Safer Human-Robot InteractionAssessing
robotic safety using the Head Injury Criterion
(HIC) and the Chamois Laceration Scale (CLS)

• Samson Phan

2
Motivation
S. Oberer, 2006
3
Outline

• Background
• Injuries
• Injury Criteria
• Test setup
• Robot platforms
• Targets
• End effectors
• Results
• Model
• Model Results
• Conclusion

4
Injuries
Bruise Penetrating Wound Burns Fractures Join
t dislocation Concussion Sprain Shock Amputati
on
S. Haddadin, 2006
5
Injuries
Bruise Penetrating Wound Burns Fractures Joint
dislocation Concussion Sprain Shock Amputatio
n
Chamois Laceration Scale (CLS)
Jettner et al. 1986
6
Factors in penetrating injuries

• Inertia
• Velocity
• Edge sharpness
• Tissue structure/ target

S. Haddadin, 2008
7
Test Setup
8
Robot Comparison

• FANUC 200iC

• Barrett WAM

9
Target

• Anthropomorphic Test Device (ATD)

• Cylindrical Headform

10
End Effectors

• Common edges in assembly plants
• Philip screwdriver point
• Stamped sheet metal edge
• Extreme edges
• 1/32 Aluminum sheet with rounded edge
• Razor blade

11
(No Transcript)
12
ResultsCLS
13
(No Transcript)
14
Coordinate Frames
z
x
x
15
(No Transcript)
16
ResultsHIC
17
Results Comparison of Injury Indices
18
Results Summary
Injury Criteria Inertia End Effector Speed Target
CLS No effect Big effect Small effect No effect
HIC Big effect No effect Big Effect Big Effect
No Effect Small Effect Large Effect

19
Modelling
Property Value
Mhead 4.5 kg
Mskin 0.06 kg
Kneck 18.75 N/m
Kskin 1x107 N/m2
Mskin 0.06 kg
Cskin 10 N-s/m
Herbst, 1998 Kappon, 2001 Sulzer ,2006
20
(No Transcript)
21
Model Real Data Comparison
22
(No Transcript)
23
Conclusions

• CLS number is highly end effector dependent
• HIC number depends on speed, inertia and target
  type
• Impact event can be modeled using a lump
  parameter system
• HIC number difference between 2 different targets
  can be explained by neck constraint
• CLS and HIC test for 2 different injury
  modalities

24
Future Work

• Refine model and apply to future designs
• Payload specific safety rating and control
• Alternative injury models

25
References

• H. Cappon et al. 2001 Development and Evaluation
  of a new Rear-Impact Crash Dummy the RID2 45th
  Stapp Car Crash Conference
• B. Herbst et al.1998 Fidelity of Anthropometric
  Test Dummy Necks in Rollover Accidents 16th
  Annual Enhanced Safety of Vehicles Conference
• S. Haddain et al. 2007 Safety Evaluation of
  Physical Human-Robot Interaction via
  Crash-Testing Proceedings of the 2007 Robotics
  Science and Systems
• S. Haddain et al. 2008 Soft-tissue Injury Caused
  by Sharp Tools Definitions, Experiments and
  Countermeasures IEEE Robotics Automation
  Magazine
• J. Jarvinen et al.1995 Analysis of self-reported
  accidents attributed to advanced manufacturing
  systems. International Journal of Human Factors
  in Manufacturing
• S. Oberer et al. 2006 Investigation of
  Human-Robot Impact International Symposium on
  Robotics
• J. Sulzer et al. 2006 Simplified MADYMO Model of
  the IHRA Head-form Impactor SAE Technical Paper
  2006-01-2349
• U. A. Workers, Review of Robot Injuries - One of
  the Best Kept Secrets, UAW Health and Safety
  Department Publication Nr.248.

26
Thank you!!

• Labmates (Barrett Hyneman, Alexis, Desbiens, Paul
  Day, Dan Aukes, Noe Esparza, Aaron Parness, John
  Ulmen, Alan Asbeck and Dan Santos)
• Professors (Mark Cutkosky, Dave Beach, Mark
  Levenston, Scott Delp, Oussama Khatib, Charles
  Taylor )

• GM R D (Charles Wampler, Javier Alcazar, Dalong
  Gao, Leandro Barajas, Jane Shi, Suzanne Hoffmann,
  Jim Wells)
• Friends (Steven Gao, Heather McCoullough, James
  Mack, Hanmao Hai, Nick Lee, Amanda Villalobos )

27
End back up slides
28
The very real danger

• Kenji Urada
• UAW Publication 248
• Jari Jarvinen (1995)
• OSHA Directive TED 01-00-015
• ISO 10218-1 (2006)

Morbidity and Mortality Weekly Report of the CDC
, 1985
29
CLS
Degree Outer Chamois Inner Chamois Rubber Dummy Face
0 None None None
1-Minimal Abrasions. Cuts to ¾-none through None None
2-Minor Abrasions. Cuts over ¾-none through None None
3-Minor As in 2 above, but one ¾ cut through Abrasions None
4-Moderate Two or three ¾ cuts through Cuts, but not through None
5-Moderate Unlimited Cuts Only one ¾ cut through None
6-Severe Unlimited Cuts Two or three cuts through to 1 ½ None
7-Severe Unlimited Cuts Unlimited Cuts Abrasions
8-Severe Unlimited Cuts Unlimited Cuts Cuts up to 1/32 deep and ¾ long
9-Very Severe Unlimited Cuts Unlimited Cuts One cut longer deeper than 8
10-Very Severe Unlimited Cuts Unlimited Cuts Numerous cuts worse than 9
30
(No Transcript)
31
(No Transcript)
32
Why the difference?
33
Results chamois effect on HIC
1.5 m/s Mean Std Dev N
ChamoisN 4.9258 0.69911 8
ChamoisY 3.546 1.00874 8
Difference 1.3798 16
2.2 m/s Mean Std Dev N
ChamoisN 6.0119 0.1 2
ChamoisY 5.4913 0.277 3
Difference 0.5206 5
34
Simplification of dummy
35
Simplification of robot manipulator
36
ResultsHIC
37
Corning Scale
Degree Outer Chamois Inner Chamois Rubber Dummy Face
0 None None None
1-Minimal Abrasions. Cuts to ¾-none through None None
2-Minor Abrasions. Cuts over ¾-none through None None
3-Minor As in 2 above, but one ¾ cut through Abrasions None
4-Moderate Two or three ¾ cuts through Cuts, but not through None
5-Moderate Unlimited Cuts Only one ¾ cut through None
6-Severe Unlimited Cuts Two or three cuts through to 1 ½ None
7-Severe Unlimited Cuts Unlimited Cuts Abrasions
8-Severe Unlimited Cuts Unlimited Cuts Cuts up to 1/32 deep and ¾ long
9-Very Severe Unlimited Cuts Unlimited Cuts One cut longer deeper than 8
10-Very Severe Unlimited Cuts Unlimited Cuts Numerous cuts worse than 9
38
Head Injury Criterion
McHenry, 2004 HIC and the ATB

• Based on kinematic data only
• No direct demonstration of functional brain
  damage
• Resultant translational acceleration produces
  pressure gardients-gt shear strain induced injury

39
(No Transcript)
40
Cylindrical Headform

• Head structure of ATD is a padded rigid aluminum
  shell
• A hemispherical headform is used in safety
  testing.