Design Method of 3D Printable Ergonomically Personalized Stabilizer - HCII 2019 presentation

Published on Aug 8, 2019

This slide was presented at HCI International 2019 Oral Presentation "Human modeling for nursing and medical art".

1. Design Method of 3D-Printable Ergonomically Personalized Stabilizer Ryota Kawamura1, Kazuki Takazawa1, Kenta Yamamoto1, Yoichi Ochiai1,2 1University of Tsukuba, 2Pixie Dust Technologies, Inc
2. 2 BACKGROUND
3. 3 Why We Need Stabilizer Blurry Video No Stabilizer, No Unblurry Video Shooting with Moving Target
4. 4 Image Stabilization How to realize Stabilization Optical Method Electronic Method [Oshima, Mitsuaki, et al. 1989] [Liu, Shuaicheng, et al. 2014] [Liu, Shuaicheng, et al. 2013]
5. 5 Gimbal Stabilizer • Need Technical Skills • Heavy • Expensive Hand-held Attached to the user How to realize Stabilization
6. 6 Personalized Stabilizer Hand-held Attached to the user Mass Produced Order Maid Personalized Ours History of Manufacturing
7. 7 Study01: Exo-Balancer
8. 8 Assembly Process (i) 3D Printing (ii) Assembling Parts (iii) Attaching Stabilizer How to carry out our method
9. A B C F F n l L x θa θb camera mc mf 9 Force Relationship • Ask the user for the position of natural holding • Under the condition of F=F’ • Derive x and l Force Relationship Equations How to carry out our method
10. 10 Experimental Protocol How Effective Our Method is Participants 8 people; age: 18~25 Experimental Procedure Continue to shoot for 30s the sphere moving within the projected area
11. 11 Experimental Protocol Experimental method list A) Proposed Method B) Proposed Method +50mm (x) C) Proposed Method -50mm (x) H) Without Proposed Method B and C are given error values to the proposed method.
 H is not using the proposed method. A B C F F n l L x θa θb camera mc mf How Effective Our Method is
12. 12 Quantitative Evaluation Acceleration sensor fixed to the camera Start End Trajectory of a sphere projected on a wall iPhone6 as an Acceleration Sensor Comparing camera shake by acceleration value How Effective Our Method is
13. 13 !"# !"$ !"% & &"& &"' !"#$% !"#&% !"#'% !"# !"$ !"% !"#$% !"#&% !"#'% ! !"# !"# !"$ !"% & &"& &"' !"#$% !"#&% !"#'% !"# !"$ !"% !"#$% !"#&% !"#'% ! !"# !"# !"$ !"% & &"& &"' !"#$% !"#&% !"#'% !"# !"$ !"% !"#$% !"#&% !"#'% ! !"# 0 30.00 30.00 30.0Time [sec] Time [sec] Time [sec] Acceleration[G]Acceleration[G]Acceleration[G] A-B A-B A-B A-C A-H A-H A-H A-C A-C XXX Y Z Z Z Y Y Quantitative Evaluation A-B: 
 No Noticeable Difference A-C: 
 C reduced the camera shake How Effective Our Method is Experimental method list A) Proposed Method B) Proposed Method +50mm (x) C) Proposed Method -50mm (x) H) Without Proposed Method
14. 14 !"# !"$ !"% & &"& &"' !"#$% !"#&% !"#'% !"# !"$ !"% !"#$% !"#&% !"#'% ! !"# 0 30.030.0me [sec] Time [sec] A-H A-H A-H XX Z Z Y Y Quantitative Evaluation A-H: 
 Camera shake is smaller in H How Effective Our Method is Experimental method list A) Proposed Method B) Proposed Method +50mm (x) C) Proposed Method -50mm (x) H) Without Proposed Method
15. 15 Qualitative Evaluation Q1: This method is easy to take photos. Q2: You feel this method is stable. Q3: You feel tired when using this shooting method. Q4: This method is designed to personalize your body. Q5: How enjoyable it this method? Q6: How stressed when you are took in this way? Question List (Likert Scale) How Effective Our Method is
16. 16 Q1 Q2 Q3 1 2 3 4 5 Q4 Q5 Q6 This method is easy to take photos. You feel this method is stable. You feel tired when using this shooting method. This method is designed to personalize your body. How enjoyable it this method?. How stressed when you are took in this way? ! " # $ ! " # $ ! " # $ ! " # $ ! " # $ ! " # $ SDM Strongly disagree Neutral Strongly agree H A B C H A B C H A B C H A B C H A B C H A B C Wilcoxon signed rank test was performed. Statistically significant difference was only found between C and H in Q6. Self-attribution tends to be higher if fixed as in the proposed method. Qualitative Evaluation How Effective Our Method is
17. 17 Study01: Summary No Stabilizer, No Unblurry Video Quantitative It is less likely to blur on hand than this method. Qualitative Self-attribution is enhanced by the fixation mechanism. Positive response for personalization with this method.
18. 18 Study01: Problem No Stabilizer, No Unblurry Video 1.Low durability because of the property of ready-made goods 2.Low adaptability for physically characterized people 3.Low stabilizing ability
19. 19 Study01: Problem No Stabilizer, No Unblurry Video 1.Low durability because of the property of ready-made goods 2.Low adaptability for physically characterized people 3.Low stabilizing ability →All 3D Printed →Modify Optimization Method →Update Stabilization Mechanism
20. 20 Study02: Topological Approach
21. 21 Structure Optimization Next Optimization Approach Our next
 approach Structure Optimization
 (Topology Optimization) Demension Optimization (Changing Parameters) Shape Optimization
 (Changing Border) Our first
 approach
22. 22 Re-Design Process camera Define Parameters along Force Equations Topology Optimization Pitch Directional Stabilization First Approach Advanced Approach 3D Printing Assemble Ready-Made Parts Renew Our Design Process All 3D Printed
 Parts 3D Printed Objects
 with Ready-Made Parts
23. 23 Topological Optimization How to optimize Topology Base Form 4 Supporting Points Loaded Surface Result and Attachment
24. 24 Topological Optimization Perspective View Right View Iteration n = 10 n = 70n = 50n = 30 Results of Topology Optimization Shape Results after Iterations
25. 25 Stabilization Accessories Pitch Directional Stabilizer for Compact Camera Updating Stabilization Mechanism
26. 26 Discussion For Next Research 3D Model after Topology Optimization with laplacian smoothing
27. Thank You Ryota Kawamura1, Kazuki Takazawa1, Kenta Yamamoto1, Yoichi Ochiai1,2 1University of Tsukuba, 2Pixie Dust Technologies, Inc Design Method of 3D-Printable Ergonomically Personalized Stabilizer

Transcript

1. Design Method of 3D-Printable Ergonomically Personalized Stabilizer Ryota Kawamura1, Kazuki Takazawa1, Kenta Yamamoto1, Yoichi Ochiai1,2 1University of Tsukuba, 2Pixie Dust Technologies, Inc
2. 2 BACKGROUND
3. 3 Why We Need Stabilizer Blurry Video No Stabilizer, No Unblurry Video Shooting with Moving Target
4. 4 Image Stabilization How to realize Stabilization Optical Method Electronic Method [Oshima, Mitsuaki, et al. 1989] [Liu, Shuaicheng, et al. 2014] [Liu, Shuaicheng, et al. 2013]
5. 5 Gimbal Stabilizer • Need Technical Skills • Heavy • Expensive Hand-held Attached to the user How to realize Stabilization
6. 6 Personalized Stabilizer Hand-held Attached to the user Mass Produced Order Maid Personalized Ours History of Manufacturing
7. 7 Study01: Exo-Balancer
8. 8 Assembly Process (i) 3D Printing (ii) Assembling Parts (iii) Attaching Stabilizer How to carry out our method
9. " # $ '` ' O M - Y ВB ВC DBNFSB ND NG 9 Force Relationship • Ask the user for the position of natural holding • Under the condition of F=F’ • Derive x and l Force Relationship Equations How to carry out our method
10. 10 Experimental Protocol How Effective Our Method is Participants 8 people; age: 18~25 Experimental Procedure Continue to shoot for 30s the sphere moving within the projected area
11. 11 Experimental Protocol Experimental method list A) Proposed Method B) Proposed Method +50mm (x) C) Proposed Method -50mm (x) H) Without Proposed Method B and C are given error values to the proposed method.
 H is not using the proposed method. " # $ '` ' O M - Y ВB ВC DBNFSB ND NG How Effective Our Method is
12. 12 Quantitative Evaluation Acceleration sensor fixed to the camera Start
    End Trajectory of a sphere projected on a wall iPhone6 as an Acceleration Sensor Comparing camera shake by acceleration value How Effective Our Method is
13. 13 !"# !"$ !"% & &"& &"' !"#$% !"#&% !"#'% !"# !"$ !"% !"#$% !"#&% !"#'% ! !"# !"# !"$ !"% & &"& &"' !"#$% !"#&% !"#'% !"# !"$ !"% !"#$% !"#&% !"#'% ! !"# !"# !"$ !"% & &"& &"' !"#$% !"#&% !"#'% !"# !"$ !"% !"#$% !"#&% !"#'% ! !"#       Time [sec] Time [sec] Time [sec] Acceleration [G] Acceleration [G] Acceleration [G] A-B A-B A-B A-C A-H A-H A-H A-C A-C X X X Y Z Z Z : Y Quantitative Evaluation A-B: 
 No Noticeable Difference A-C: 
 C reduced the camera shake How Effective Our Method is Experimental method list A) Proposed Method B) Proposed Method +50mm (x) C) Proposed Method -50mm (x) H) Without Proposed Method
14. 14 !"# !"$ !"% & &"& &"' !"#$% !"#&% !"#'% !"# !"$ !"% !"#$% !"#&% !"#'% ! !"#    me [sec] Time [sec] A-H A-H A-H X X Z Z : Y Quantitative Evaluation A-H: 
 Camera shake is smaller in H How Effective Our Method is Experimental method list A) Proposed Method B) Proposed Method +50mm (x) C) Proposed Method -50mm (x) H) Without Proposed Method
15. 15 Qualitative Evaluation Q1: This method is easy to take photos. Q2: You feel this method is stable. Q3: You feel tired when using this shooting method. Q4: This method is designed to personalize your body. Q5: How enjoyable it this method? Q6: How stressed when you are took in this way? Question List (Likert Scale) How Effective Our Method is
16. 16 Q1 Q2 Q3      Q4 Q5 Q6 This method is easy to take photos. You feel this method is stable. You feel tired when using this shooting method This method is designed to personalize your body. How enjoyable it this method?. How stressed when you are took in this way? ! " # $ ! " # $ ! " # $ ! " # $ ! " # $ ! " # $ SD M Strongly disagree Neutral Strongly agree H A B C H A B C H A B C H A B C H A B C H A B C Wilcoxon signed rank test was performed. Statistically significant difference was only found between C and H in Q6. Self-attribution tends to be higher if fixed as in the proposed method. Qualitative Evaluation How Effective Our Method is
17. 17 Study01: Summary No Stabilizer, No Unblurry Video Quantitative It is less likely to blur on hand than this method. Qualitative Self-attribution is enhanced by the fixation mechanism. Positive response for personalization with this method.
18. 18 Study01: Problem No Stabilizer, No Unblurry Video 1.Low durability because of the property of ready-made goods 2.Low adaptability for physically characterized people 3.Low stabilizing ability
19. 19 Study01: Problem No Stabilizer, No Unblurry Video 1.Low durability because of the property of ready-made goods 2.Low adaptability for physically characterized people 3.Low stabilizing ability →All 3D Printed →Modify Optimization Method →Update Stabilization Mechanism
20. 20 Study02: Topological Approach
21. 21 Structure Optimization Next Optimization Approach Our next
 approach Structure Optimization
 (Topology Optimization) Demension Optimization (Changing Parameters) Shape Optimization
 (Changing Border) Our first
 approach
22. 22 Re-Design Process DBNFSB %FpOF
    1BSBNFUFST BMPOH
    'PSDF
    &RVBUJPOT 5PQPMPHZ
    0QUJNJ[BUJPO 1JUDI
    %JSFDUJPOBM
    4UBCJMJ[BUJPO 'JSTU
    "QQSPBDI "EWBODFE
    "QQSPBDI %
    1SJOUJOH "TTFNCMF 3FBEZ.BEF
    1BSUT Renew Our Design Process All 3D Printed
 Parts 3D Printed Objects
 with Ready-Made Parts
23. 23 Topological Optimization How to optimize Topology Base Form 4 Supporting Points Loaded Surface Result and Attachment
24. 24 Topological Optimization 1FSTQFDUJWF 7JFX 3JHIU 7JFX *UFSBUJPO O
    
     O
    
     O
    
     O
    
     Results of Topology Optimization Shape Results after Iterations
25. 25 Stabilization Accessories Pitch Directional Stabilizer for Compact Camera Updating Stabilization Mechanism
26. 26 Discussion For Next Research 3D Model after Topology Optimization with laplacian smoothing
27. Thank You Ryota Kawamura1, Kazuki Takazawa1, Kenta Yamamoto1, Yoichi Ochiai1,2 1University of Tsukuba, 2Pixie Dust Technologies, Inc Design Method of 3D-Printable Ergonomically Personalized Stabilizer