1 / 55

CAGING OF RIGID POLYTOPES VIA DISPERSION CONTROL OF POINT FINGERS

CAGING OF RIGID POLYTOPES VIA DISPERSION CONTROL OF POINT FINGERS Peam Pipattanasomporn Advisor: Attawith Sudsang. Motivation?. !. !. !. Better Approach?. Overview. Proposed Ph.d . Thesis. Additional Chapters. Master Thesis. 2-Squeeze (2006). n-Squeeze (2008). L. X. H.

lizina
Télécharger la présentation

CAGING OF RIGID POLYTOPES VIA DISPERSION CONTROL OF POINT FINGERS

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. CAGING OF RIGID POLYTOPES VIA DISPERSION CONTROL OF POINT FINGERS PeamPipattanasompornAdvisor: AttawithSudsang

  2. Motivation? ! ! !

  3. Better Approach?

  4. Overview Proposed Ph.d. Thesis Additional Chapters Master Thesis 2-Squeeze(2006) n-Squeeze(2008) L X H Fix Cage(2011) S C Robust Cage(2012) 2-Stretch(2006) n-Stretch(2008) Imperfect Shape (2010)

  5. Overview Proposed Ph.d. Thesis Additional Chapters Master Thesis 2-Squeeze(2006) n-Squeeze(2008) L X H Fix Cage(2011) S C Robust Cage(2012) 2-Stretch(2006) n-Stretch(2008) Imperfect Shape (2010)

  6. 2-Squeeze, How? • Keep distance below a value • Given object shape, solve: • Where to place the fingers? • The upperbounddistance? H “Distance”

  7. 2-Squeeze • Possible escape path (object frame) H Distance Along the path

  8. 2-Squeeze • “Better” escape path H Distance Upperbound “Better” Initial Along the path

  9. 2-Squeeze • Find an Optimal Escape Path in C-Free H (abstracted)C-Obstacle Abstracted set ofescape configurations a (a,b) b Configuration Space (4D) Workspace (2D)

  10. 2-Squeeze • Find an Optimal Escape Path in C-Free (abstract)C-Obstacle Abstract set ofescape configurations (a,b) Configuration Space (4D)

  11. C-Free Decomposition C-Obstacle

  12. Paths connecting Terminals C-Obstacle

  13. Finite Categorization of Paths C-Obstacle

  14. Straight Path Distance : |a-b|2 a b (linear interpolation) Along the path (a,b)

  15. Moving Across Convex Subsets C-Obstacle

  16. Through Convex Intersections C-Obstacle

  17. Requirements For The Algorithm Distance(x) & x Convex Rigid Transformation Invariant

  18. Convex & RTI Examples • d1+ d2+ d3 • d12+ d22+ d32 • max(d1, d2, d3) x1 d3 d1 x3 x2 d2 • Larger  Loose cage • Fingers at a point  Smallest “Formation Size”

  19. Results (n-Squeeze) Size: d12+d22+d32+d42

  20. Squeezing? 1 1 1 1 3 3 2 2 3 2 3 2 1-DOF Scaling ONLY

  21. “Size” & “Deformation” Smaller sizeSlightly Deformed 1 1 1 1 3 2 3 2 3 2 2 3 Reference Formation Same sizeNo deformation Larger sizeDeformed

  22. “Size” & “Deformation” Smaller sizeSlightly Deformed 1 1 1 1 3 2 3 3 2 2 2 3 Reference Formation Same sizeNo deformationSame Formation Larger sizeDeformed

  23. “Size” & “Deformation” Smaller sizeSlightly Deformed 1 1 1 1 3 2 3 3 2 2 2 3 Reference Formation Same sizeNo deformationSame Formation Larger sizeDeformed

  24. “Size” & “Deformation” Smaller sizeSlightly Deformed 1 1 1 1 3 2 3 3 2 2 2 3 Reference Formation Same sizeNo deformationSame Formation Larger sizeDeformed

  25. “Size” & “Deformation” • |r|22(x) = |A†x|22 • “Scale” or “Size” (w.r.t. reference) • D(x) = |A(r;t) – x|22 • “Deformation upto Scale” (w.r.t. reference) 1 1 3 2 2 3 A stores information of the reference.

  26. Squeezing ? Convex& RTI Size = |r|22<??? D ≤ 0 & 1 1 1 3 2 3 2 3 2 Convex& RTI 1-DOF Scaling ONLY

  27. Squeezing |r|22; D ≤ 0  ; D > 0 Size* = Size = |r|22 <??? D ≤ 0  &  1 1 1 D ≤ 0 D > 0 D > 0 3 2 3 2 3 2 x 1-DOF Scaling ONLY

  28. Fix Formation Cage Convex& RTI ConvexConstraint Size* = 1 1 Size* ≤ 1 Size*  1 3 2 & “Squeeze” “Stretch”

  29. Robust Caging • Keep error (deformation) below a value • Given object shape, find: • Where to place the fingers • The upperbounderror Independent Capture Regions

  30. n-Squeeze vs Fix Formation X KEEP SIZE ERROR (DEFORMATION) BELOW UPPERBOUND BELOW UPPERBOUNDOPTIMAL ESCAPE PATH SIZE MINIMIZE UPPERBOUND DISTANCE ERROR (DEFORMATION)

  31. Error Tolerance  1 1 1 1 • r+t- inf D2 =  2 2 2 2 r,tϵR2  3 3 3 3 2 “Placement Error upto Scale” “Placement Error”  1 1 1 1 inf Ep = • r+t-  2 2 2 2 |r|2=1tϵR2  3 3 3 3 p NOT CONVEX!

  32. Approximation  infg(r)r ϵRi infg(r)|r|2=1  mini ϵ{1,…, m} R2 R3 R1 R4

  33. Approximation  infg(r)r ϵRi infg(r)|r|2=1  mini ϵ{1,…, m} R2 Min of Convex Functions(not convex) R3 R1 R4

  34. Optimal Path Min of a Convex Function is Convexf = f1 = min(f1)

  35. Optimal Path f1 = f f = f2 f1 = f = f2 Min of Two Convex Functionsf = min(f1, f2) 35

  36. Optimal Path f1 = f f = f2 f1 = f = f2 ??? Min of Two Convex Functionsf = min(f1, f2) 36

  37. Optimal Path What is the optimal path, starting from the minimal points? f1=f=f2 f(x) f=f1 f=f2 2 1 x

  38. Critical Point Consider… f1=f=f2 f(x) f=f1 f=f2 1,2 2 1 x Only the points under the water level are reachable when the maximum deformation is limited to below the water level.

  39. Optimal Path : minimizer for a CONVEX optimization problem:minimize L s.t.f1(x) < Lf2(x) < L f1=f=f2 f(x) f=f1 f=f2 1,2 Critical Value 2 1 x

  40. Critical Point f1=f=f2 f(x) f=f1 f=f2 1,2 Critical Value 2 1 x

  41. Min of Multiple Convex Functions f= f2 f= f1 f= f3 Min of Multiple Convex Functionsf = min(f1, f2 , f3)

  42. Min of Multiple Convex Functions 2,3 f= f2 1,2 2 3 f= f1 f= f3 1,3 1 Min of Multiple Convex Functionsf = min(f1, f2 , f3)

  43. Search Space 2 1,2 1,3 1 3 2,3 Min of Multiple Convex Functionsf = min(f1, f2 , f3) Include all possible between any two regions: f=fi, f=fj

  44. Optimal Path

  45. Results

  46. Results

  47. Shape Uncertainty Scanned Object Exact Object(Unknown) sensor

  48. Idea • Cage subobject Cage object ? • Fingers must not penetrate the object. H

  49. Idea • Find placements that cage subobject, outside superobject. Exact Object(Unknown) Exact boundary (unknown) but inbetween the bounds.

  50. Applications • Simplification • Curved Surface, Spherical Fingers • Shape Uncertainty • Slightly Deformable Object • Partial Observation

More Related