SLIDE 1
Placing Novel Objects
Robotic grasping has been studied for decades.
Example scenarios:
Arranging grocery in a fridge. Loading a dish-rack. Organizing a house.
2 Jiang, Zheng, Lim and Saxena
Learning to Place New Objects Yun Jiang , Changxi Zheng, Marcus Lim - - PowerPoint PPT Presentation
Learning to Place New Objects Yun Jiang , Changxi Zheng, Marcus Lim - - PowerPoint PPT Presentation
Learning to Place New Objects Yun Jiang , Changxi Zheng, Marcus Lim and Ashutosh Saxena Cornell University Placing Novel Objects Robotic grasping has been studied for decades. How to place objects after grasping? Example scenarios:
SLIDE 2
SLIDE 3
Placing Problem Formulation
Input: point cloud of objects and placing areas (from Kinect). Output: a stable and preferred placement specified by 3D
location/orientation of the object
3 Jiang, Zheng, Lim and Saxena
SLIDE 4
Challenges
Where to place?
complex (non-flat) areas Semantic preferences: shoes not on
tables
Which orientation to place in?
Depends on different placing areas.
4 Jiang, Zheng, Lim and Saxena
SLIDE 5
Previous Work
Where to place?
Flat and clutter-free surface
[M.J.Schuster et al., 2010]
How to execute placing?
Given desired location to place, Path planning [Lozano-Perez et al., 2002]. Tactile feedback [Edsinger and Kemp, 2002].
A.Edsinger and C.C.Kemp, 2002 M.J.Schuster et al., 2010 5 Jiang, Zheng, Lim and Saxena
SLIDE 6
Learning Approach
Placing single object
6 Jiang, Zheng, Lim and Saxena
SLIDE 7
Stability Features
Supporting contacts (12)
Falling distance, Eigen values,
center of mass, etc.
7 Jiang, Zheng, Lim and Saxena
contacts
SLIDE 8
Stability Features
Supporting contacts (12) Caging features (37)
Height Horizontal distance from the placing area to the object
8 Jiang, Zheng, Lim and Saxena
(b) top view (a) side view
SLIDE 9
Semantic Features
Supporting contacts (12) Caging features (37) Histogram features (96)
# points from object, placing area and their ratio
9 Jiang, Zheng, Lim and Saxena
(b) top view (a) side view
SLIDE 10
Max-margin learning
Jiang, Zheng, Lim and Saxena 10
Features for each placement Labels Soft-margin support vector machine (SVM) Shared-sparsity in the parameters.
SLIDE 11
Learning Approach
11 Jiang, Zheng, Lim and Saxena
SLIDE 12
Learning Experiments
Jiang, Zheng, Lim and Saxena 12
Finding best location and orientation
7 placing areas and 19 objects A total of 620 placements
SLIDE 13
Robotic Experiments
13 Jiang, Zheng, Lim and Saxena
SLIDE 14
Robotic Experiments
Jiang, Zheng, Lim and Saxena 14
Objects/environments are never seen before by the robot!
T
- tal 400 robotic trials.
SLIDE 15
Placing multiple objects
Jiang, Zheng, Lim and Saxena 15
Task: placing multiple objects into multiple placing areas Challenges
(1) stability (2) semantic preference (3) linear stacking (4) non-overlap
Max-margin learning Inference as a linear programming problem [ Jiang, Zheng, Lim and Saxena, IJRR 2012 ]
SLIDE 16
Experiments: Full Scene Semantics
Jiang, Zheng, Lim and Saxena 16
Finding semantically preferred placing area
98 objects from 16 categories 11 different placing areas
SLIDE 17
Video
Jiang, Zheng, Lim and Saxena 17
SLIDE 18
Learning Object Arrangements using Human Context
Jiang, Zheng, Lim and Saxena 18
Learn human-object relationships Arrange objects meaningfully
Jiang, Lim and Saxena, ICML’12
SLIDE 19
Thank you!
Jiang, Zheng, Lim and Saxena 19
Learning to Place New Objects.
Jiang, Zheng, Lim and Saxena, ICRA’12.
Learning to Place New Objects in a Scene.
Jiang, Zheng, Lim and Saxena, IJRR’12
Learning Object Arrangements in 3D Scenes using Human Context.
Jiang, Lim and Saxena, ICML’12
Dataset for object placing & arrangements
98 objects from 16 categories, 40 placing areas 180 human labeled arrangements on 20 rooms and 47 objects