FORGE: Force-Guided Exploration for Robust Contact-Rich Manipulation Under Uncertainty
- 14 March 2025
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 10 (5) , 4436-4443
Abstract
We present FORGE, a method for sim-to-real transfer of force-aware manipulation policies in the presence of significant pose uncertainty. During simulation-based policy learning, FORGE combines a force threshold mechanism with a dynamics randomization scheme to enable robust transfer of the learned policies to the real robot. At deployment, FORGE policies, conditioned on a maximum allowable force, adaptively perform contact-rich tasks while avoiding aggressive and unsafe behaviour, regardless of the controller gains. Additionally, FORGE policies predict task success, enabling efficient termination and autonomous tuning of the force threshold. We show that FORGE can be used to learn a variety of robust contact-rich policies, including the forceful insertion of snap-fit connectors. We further demonstrate the multistage assembly of a planetary gear system, which requires success across three assembly tasks: nut threading, insertion, and gear meshing.Keywords
Funding Information
- NVIDIA and DEVCOM Army Research Laboratory (W911NF-23-2-0012)
This publication has 32 references indexed in Scilit:
- DeXtreme: Transfer of Agile In-hand Manipulation from Simulation to Reality2023
- Towards Real-World Force-Sensitive Robotic Assembly through Deep Reinforcement Learning in Simulations2021
- Learning Sequences of Manipulation Primitives for Robotic Assembly2021
- Robotic assembly of timber joints using reinforcement learningAutomation in Construction, 2021
- Sim-to-Real Transfer of Robotic Control with Dynamics Randomization2018
- Learning Manipulation Graphs from Demonstrations Using Multimodal Sensory Signals2018
- Supersizing self-supervision: Learning to grasp from 50K tries and 700 robot hours2016
- Failure detection in assembly: Force signature analysis2010
- Planning and acting in partially observable stochastic domainsArtificial Intelligence, 1998
- System Identification1998