FORGE: Force-Guided Exploration for Robust Contact-Rich Manipulation under Uncertainty

TL;DR

FORGE tackles sim-to-real transfer for force-aware manipulation under pose uncertainty in contact-rich assembly tasks.It introduces a force-threshold conditioned policy, dynamics randomization, and a success-prediction module to modulate contact forces and terminate efficiently, formalized through a POMDP with states like $p^{ee}, p^{fixed}, p^{held} \in SE(3)$ and forces $F^{ee} \in \mathbb{R}^3$.The method demonstrates robust sim-to-real transfer on peg insertion, gear meshing, and nut threading, and scales to multi-stage planetary gearbox assembly, with real-world trials exceeding 1000 runs.These contributions reduce risk of damage, improve cycle time, and enable automatic tuning of interaction force via $F_{th}$ and $p_{term}$, potentially impacting industrial manipulation tasks.

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. Project website can be accessed at https://noseworm.github.io/forge/.

Figures (10)

• Figure 1: FORGE uses force feedback to learn search behaviours for contact-rich tasks with position estimation uncertainty. It combines dynamics randomization, a force threshold, and success prediction for robust sim-to-real transfer. The resulting policies are safe and efficient (bottom) compared to aggressive baseline policies that cause parts to slip (top).
• Figure 2: FORGE is evaluated on three tasks from Factorynarang2022factory: Peg Insertion, Gear Meshing, and Nut Threading. Each task is trained solely in simulation (top) and transferred directly to the real robot (bottom).
• Figure 3: Perception Error For each task, we visualize what the different position estimation errors look like overlaid on the fixed part.
• Figure 4: Noise Analysis Performance broken down by level of position error. Each subplot is a planar representation of the error levels where each ring corresponds to low (0-1mm), medium (1-2.5mm), and high (2.5-5mm) error. Success rate, stated in black text, is also represented by the shade of the corresponding ring. Dots represent x-y noise samples for successful (green) and failed (red) trials. FORGE results in good performance across tasks even with high error levels.
• Figure 5: Gains Analysis (180 trials, 8mm Peg) With force sensing, FORGE can achieve robust success rates (bottom) across varying controller gains at deployment time. Even with different gains, force sensing allows the policy to modulate its actions to achieve low contact forces (top).