Human-in-the-loop Design Framework Dr. Onan Demirel B.S. Ph.D. - - PowerPoint PPT Presentation

▶
human in the loop design framework
SMART_READER_LITE
LIVE PREVIEW

Human-in-the-loop Design Framework Dr. Onan Demirel B.S. Ph.D. - - PowerPoint PPT Presentation

Sep 18, 2022 152 likes •607 views

School of Mechanical, Industrial, COLLEGE OF ENGINEERING and Manufacturing Engineering Human-in-the-loop Design Framework Dr. Onan Demirel B.S. Ph.D. M.S. Gabriel Kemling Salman Ahmed Kamolnat Tabattanon Valerie Byxbe Karina Roundtree

slide-1
SLIDE 1

COLLEGE OF ENGINEERING

School of Mechanical, Industrial, and Manufacturing Engineering Ph.D. Salman Ahmed Karina Roundtree Nicolás S. Zurita (co-advised with Irem Tumer) Lukman Irshad (co-advised with Irem Tumer) M.S. Kamolnat Tabattanon Alex Jennings Jianfu Zhang MihirSunil Gawand

Human-in-the-loop Design Framework

B.S. Gabriel Kemling Valerie Byxbe Timothy Slama

  • Dr. Onan Demirel
slide-2
SLIDE 2

ABOUT ME

Onan Demirel Assistant Professor School of Mechanical, Industrial, and Manufacturing Engineering Oregon State University Education: BS, MS, and PhD (2015) in Purdue University Research Interest: Design Theory and Methods Human Factors Engineering Digital Human Modeling Product Design and Development Contact: Email: onan.demirel@oregonstate.edu Website-1: https:// www.onandemirel.com Website-2: https://design.engr.oregonstate.edu/demirel 322 Rogers Hall | (765) 409-9419 | Corvallis, OR 97331

slide-3
SLIDE 3
slide-4
SLIDE 4

45,000 Americans die each year due to medical errors. Human error is a determining factor in 60% to 80% of industrial accidents. Ignoring human factors of work place costs about $4.6 billion per year. Human errors account for 80% of offshore accidents.

To err is human.....

Human error causes 94% of car accidents. 80% of accidents in commercial flights caused by pilot error.

slide-5
SLIDE 5

Who to blame?

Driver was distracted! Driver was not paying attention! Poor judgement at the intersection! Bad driver! ...

slide-6
SLIDE 6

To err is human, to forgive . divine.

(Alexander Pope, 1711)

slide-7
SLIDE 7

To err is human, to forgive . , design.

slide-8
SLIDE 8

Good news: most of these failures and accidents can be prevented or mitigated!

slide-9
SLIDE 9

PILLAR OBSCURATION

slide-10
SLIDE 10

Focus: Developing design methodologies to optimize human wellbeing and overall system performance.

slide-11
SLIDE 11

ENGINEERING HUMAN FACTORS INDUSTRIAL DESIGN

Biomechanics Cognition Psychology ..... Idea4on Crea4vity Prototyping ..... Simula4ons Op4miza4on Modeling .....

FORM FUNCTION HUMANS DIGITAL HUMAN MODELING

slide-12
SLIDE 12

A MIXED (HYBRID) PROTOTYPING TESTBED

  • Beyond 3D Modeling:

– Multi-physics simulations – Digital sculpting – ...

  • Not just physical prototypes:

– Virtual reality – Photorealistic rendering – ....

  • Physical and Cognitive aspects:

– Mental workload – Perception – ....

slide-13
SLIDE 13

HUMAN-IN-THE-LOOP FRAMEWORK

Digital Human Modeling (DHM): representation of the human inserted into a simulation

  • r a virtual environment to facilitate prediction of safety and/or performance.

DHM includes:

Visualization – (form) Math/science – (function)

DHM can evaluate/predict human-product interactions:

Biomechanics – (L4/L5 moments) Ergonomics – (comfort / discomfort)

slide-14
SLIDE 14

HUMAN-IN-THE-LOOP FRAMEWORK

Utilizes Digital Human Modeling research to integrate three domains: Design: 1.

Engineering – Industrial Design –

Human Factors Engineering: 2.

Physiology – Cognition –

Systems Engineering: 3.

Detail Design and Development – Functional Failure Analysis –

slide-15
SLIDE 15

DIGITAL HUMAN MODELING

Motion & Posture Feedback / Control Visualization Motion Capture Eye Tracking ….

Human Data Product Data Failure Data

Failure Trends Statistical Data Failure Modes Reliability Data ……

INPUT

(Design Data) Reach Envelope L4/L5 Compression RULA …. Binocular Vision Obscuration Zones Reflection Zones …. Risk of Failure Failure Paths Error Mitigation ….

Muscloskletal Analysis Vision Analysis Failure Analysis

OUTPUT

(Human-Product Interaction Assessment) Questionnaire Cooper Harper Test ….

“Objective” “Subjective”

Surface Model Parametric Model …. Physical Props Fixtures & Handles ….

“CAD” “Low-End Prototype”

HUMANS FORM FUNTION

HUMAN-IN-THE-LOOP FRAMEWORK

slide-16
SLIDE 16

HUMAN-IN-THE-LOOP FRAMEWORK

slide-17
SLIDE 17

HUMAN-IN-THE-LOOP FRAMEWORK

slide-18
SLIDE 18

Market Researh Modeling Engineering Prototyping Manufacturing …. Product Development Phases Cost Associated with Product Development Reduction in Cost ($) Conventional Design Process Concurrent Engineering Human-in-the-loop Framework

EARLY DESIGN – COST & TIME SAVINGS

slide-19
SLIDE 19

Understanding how we harness

  • technology to capture human:

Needs – Abilities – Limitations – Desires –

Implementing better products and

  • complex-systems that are:

Effective – Efficient – Engaging – Error tolerant – Easy to use/learn –

slide-20
SLIDE 20

Products that improve safety:

– Protective equipment – Lightweight cushions – .....

Products that improve wellbeing:

– Medical devices – Exoskeletons and prosthetics – .....

Modern product development :

– Transportation design – Fashion, textile and clothing – Advance manufacturing – Medical products – Consumer goods

HUMAN-IN-THE-LOOP FRAMEWORK

slide-21
SLIDE 21

COURSES

Computational design approach for modern product development ENGR 248: Engineering Graphics and 3D Modeling

  • solid modeling, drafting, rendering, industrial design

–

ME 611: Modern Product Development

  • with Dr. Rob Stone

computational design, physical and digital prototyping –

ME 599X: Digital Human Modeling for Design

  • new course offered in Winter 2018

New course includes computational human –

  • centric design

Industrial + Engineering Design

  • Simulations
  • Human
  • in-the-loop design

Biomechanics for Ergonomics

  • Computational Ergonomics
slide-22
SLIDE 22

1

ASSESSMENT OF TYPES OF PROTOTYPING IN HUMAN-CENTERED PRODUCT DESIGN

How the prototypes should be built to test a Human-Centered Product to save money, time and improve product quality?

  • Physical or Computational or Mixed prototype?
  • How many prototypes to built?
  • How to account for complex human-product interaction,

emergency situation?

  • Physical Prototypes
  • Computational Prototypes
slide-23
SLIDE 23

2

APPROACH & METHOD

slide-24
SLIDE 24

3

  • CAD (Computer Aided Design) is used to create Workplace
  • VR (Virtual Reality) is used to immerse a human subject in the Workplace
  • Kinect is used to capture human motion and DHM (Digital Human Modeling)

is used for ergonomic assessment

https://www.google.com/search?q=kinect+1&rlz=1C1GGRV_enUS748US748&source=lnms&tbm=isch&sa=X&ved=0ahUKEwig4oyppqHaAhXLwlQKHWgLBB0Q_AUICygC&biw=1600&bih=1109#imgrc=drpMzoO9plgekM: https://www.google.com/search?q=htc+vive&rlz=1C1GGRV_enUS748US748&source=lnms&tbm=isch&sa=X&ved=0ahUKEwjLwuPDpqHaAhULilQKHcjHC4IQ_AUIDSgE&biw=1600&bih=1109#imgdii=zR-Kza_9HNz18M:&imgrc=9iaJ9d7TKoCKgM:

APPROACH & METHOD

KINECT HTC VIBE

slide-25
SLIDE 25

4

APPROACH & METHOD

slide-26
SLIDE 26

5

Computational Prototype using CAD and DHM (JACK) Mixed Prototype using VR, Human Subject, Kinect, CAD,DHM

APPROACH & METHOD

slide-27
SLIDE 27

6

RESULTS

Reaching Task Computational Prototype (JACK) Mixed Prototype

slide-28
SLIDE 28

7

RESULT

Reaching Task Computational Prototype (JACK) Mixed Prototype

slide-29
SLIDE 29

8

15 15 16 7 13 20 27 8 5 10 15 20 25 30 1 2 3 4

Posture Analysis

Routine Emergency

Upper Arm Flexion Angle

Reach Locations

Oxygen Mask Reach Front Panel Reach Circuit Breaker Reach Throttle Reach

RESULT

slide-30
SLIDE 30

9

FUTURE WORKS

  • Use higher fidelity MoCap devices
  • Employ more subjects
  • Create higher fidelity prototypes by including smokes, fire, etc.
slide-31
SLIDE 31

Assistive and Adaptive Technology: Impacts on Human Performance for Early Design Phase Considerations

Presenter: Karina Roundtree

slide-32
SLIDE 32

Examples of Current Technology

1

Ground Collision Avoidance System (Proximity) Global Positioning System (Navigation) Automated Radar Plotting Aid (Collision Avoidance)

[1] [2] [3]

slide-33
SLIDE 33

Human Error

2

Limited Human Capabilities Stress Lack of Situational Awareness Failure of Vigilance Perceived Risk Easily Interrupted

[4]

Fatigue Negative Emotions Poor Timing Confusion

slide-34
SLIDE 34
  • Workload = amount of effort and energy an individual invests into a task +

level of work necessary to complete task

  • Affects human performance
  • Encompasses
  • Operator
  • Task
  • System Demands
  • Environment
  • Yerkes-Dodson Law
  • Performance vs. Arousal

Workload

3 [5]

slide-35
SLIDE 35

Human and Machine Interaction

4 [5]

T

  • m

u c h Too little Perfect amount

  • f automation
slide-36
SLIDE 36
  • Traditional Levels of Automation
  • Sheridan and Verplank’s 10 Levels of Automation
  • National Highway Traffic Safety Administration Classification of Vehicle

Automation

  • Pilot Authorization and Control of Tasks (PACT) Framework

Levels of Automation vs. Adaptive Automation

5

  • Adaptive Automation
  • Capable of changing levels of automation by monitoring
  • Cognitive and physical aspects of operator
  • Task Difficulty
  • Environment

[6]

slide-37
SLIDE 37
  • How do humans and machines independently and concurrently
  • Communicate
  • Understand
  • Act upon given information
  • Humans decision process
  • Human Performance
  • Subjective Measurements
  • Psychophysiological Measurements
  • Performance Measurements

Do these assistive technologies work? How do we know?

6

Visual Auditory Tactile Smell Taste Bottom-Up Vs. Top-Down Skills Rules Knowledge Senses Process Decision

slide-38
SLIDE 38

Subjective Measurements

  • Used to estimate users’ perceptions of

experiences

  • Benefits
  • Easy to implement
  • Low cost
  • Non-intrusive
  • Examples
  • Likert Scale
  • National Aeronautics and Space

Administration Task Load Index (NASA-TLX)

  • Additional Subjective Measurement Tools
  • Visual Analog Scale
  • Subjective Workload Assessment Technique
  • Rating Scale Mental Effort
  • Cooper Harper Test

7 [7]

slide-39
SLIDE 39

Psychophysiological Measurements

  • Most popular human performance assessment
  • Benefits
  • Continuous measurements
  • Not susceptible to user bias
  • Types
  • Cardiovascular
  • Skin Conductivity
  • Visual Sensory System
  • Respiration

8

Heart Period Interbeat Interval

[8] [9] [10]

slide-40
SLIDE 40

Performance Measurements

  • Number of Correct or Incorrect Responses
  • Position, Velocity, Acceleration, and Time
  • Additional Performance Measurements
  • Brain Activity
  • Facial Expressions
  • Speech Features
  • Body Movement
  • Sleep Characteristics
  • Pass or Fail Grades
  • Initiatives
  • Team Parameters

9

slide-41
SLIDE 41

Statistical Analysis

10

Research Question Data Post hoc Parametric T-test, Z-test ANOVA Pearson Correlation Non- Parametric Mann-Whitney Test, Wilcoxon Signed Rank Test Kruskal-Wallis Test, Friedman Test Spearman Correlation Chi Squared Test Regression

slide-42
SLIDE 42

Future Recommendations

11

Utilize a semi physical prototype à human performance measurements à evaluate design choices à affirm engineering specifications

[11]

slide-43
SLIDE 43

Questions?

12

slide-44
SLIDE 44

References

13

[1] “Point of Recovery,” [Online]. Available: https://www.acc.af.mil/News/Photos.aspx?igphoto=2001677419 [2]

  • G. R. J. Hockey, A. Healey, M. Crawshaw, D. G. Wastell, and J. Sauer, “Cognitive Demands of

Collision Avoidance in Simulated Ship Control,” Hum. Factors J. Hum. Factors Ergon. Soc.,

  • vol. 45, no. 2, pp. 252–265, 2003.

[3] “Garmin nuvi 40LM 4.3-Inch Portable GPS Navigator with Lifetime Maps (US),” [Online]. Available: https://www.gpsnavigatorforcarshop.blogspot.com/2013/04/garmin-nuvi-40lm- 43-inch-portable-gps.html [4] “Crazy Car Crash As Audi TT Flies Into House,” [Online]. Available: https://www.goskippy.com/2013/03/25/crazy-car-crash-as-audi-tt-flies-into-house/ [5]

  • D. M. Diamond, A. M. Campbell, C. R. Park, J. Halonen, and P. R. Zoladz, “The Temporal

Dynamics Model of Emotional Memory Processing: A Synthesis on the Neurobiological Basis

  • f Stress-Induced Amnesia, Flashbulb and Traumatic Memories, and the Yerkes-Dodson

Law,” Neural Plast., vol. 2007, pp. 1–33, 2007. [6]

  • D. Richards and A. Stedmon, “To delegate or not to delegate: A review of control

frameworks for autonomous cars,” Appl. Ergon., vol. 53, pp. 383–388, Mar. 2016. [7]

  • S. G. Hart, “NASA-Task Load Index (NASA-TLX); 20 Years Later,” in Proceedings of the Human

Factors and Ergonomics Society 50th Annual Meeting, Santa Monica, California, 2006, pp. 904–908.

slide-45
SLIDE 45

References continued

14

[8] (2009, January 15). “Circuits vs. Software,” [Online]. Available: https://www.swharden.com/wp/2009-01-15-circuits-vs-software/ [9] “NeXus Skin Conductance Sensor,” [Online]. Available: https://www.ycanaustralia.com/SKIN-CONDUCTANCE-SENSOR-FOR-NEXUS-4-10-32F [10] E. Pikaar, “Finding Alternative Ways to Measure Mental Workload While Driving,” Universiteit Leiden, 2015. [11] H. O. Demirel, “Modular Human-in-the-Loop Design Framework Based on Human Factors,” Purdue University, 2015.