Learning Transferable Push Manipulation Skills in Novel Contexts
Rhys Howard, Claudio Zito

TL;DR
This paper introduces a transferable internal model for push manipulation that enables robots to predict object movements in novel contexts by learning local contact and motion models, improving prediction accuracy with available information.
Contribution
The paper presents a novel factorized learning approach for transferable push manipulation models, combining local contact and motion models for improved prediction in new environments.
Findings
Unbiased predictor provides coarse estimates without specific environment info.
Biased predictor achieves more accurate predictions with environment-specific tuning.
Model predictions align well with physics simulator outcomes.
Abstract
This paper is concerned with learning transferable forward models for push manipulation that can be applying to novel contexts and how to improve the quality of prediction when critical information is available. We propose to learn a parametric internal model for push interactions that, similar for humans, enables a robot to predict the outcome of a physical interaction even in novel contexts. Given a desired push action, humans are capable to identify where to place their finger on a new object so to produce a predictable motion of the object. We achieve the same behaviour by factorising the learning into two parts. First, we learn a set of local contact models to represent the geometrical relations between the robot pusher, the object, and the environment. Then we learn a set of parametric local motion models to predict how these contacts change throughout a push. The set of contact…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
