Robot policies improve contact tasks by directly controlling force output
Wrench-ACT: Enhancing Robot Policies for Contact Rich Behavior Using Direct Wrench Control
Robotics
Summary
Robots that handle delicate or tricky tasks need to carefully manage how hard they push or pull on things. Most robots today decide where to move their arms but don’t directly decide how much force to use. The authors developed a new way for robots to learn by copying human operators who control force directly, which helps robots handle tasks involving physical contact better. They tested this on five different tasks and found their method worked as well or better than traditional methods based on moving to specific positions.
What this means in practice
- •For industrial robot programmers: Create robot programs that directly control forces for improving precision in tasks requiring delicate contact, such as assembly or material handling.
- •For robotics teleoperation teams: Develop training methods using force-feedback teleoperation data to improve robot control in contact-rich environments.
Authors
Johannes Hechtl, Yannik Blei, Simon Ball, Reihaneh Mirjalili, Michael Krawez, Seongjin Bien, Philipp Schmitt, Wolfram Burgard
Abstract
While contact-rich manipulation requires deliberate regulation of interaction forces, recent approaches to robot manipulation learning predominantly represent actions as target positions or poses. Even methods that incorporate force sensing either use it solely as an observation or, when predicting forces as part of the output, rely on a hybrid force controller. In this paper, we propose an imitation learning policy that predicts wrenches as its sole action output for direct use by a pure force controller. Our studies suggest that force-domain imitation learning depends critically on data collection, with force-feedback teleoperation improving policy performance by capturing the operator's deliberate force regulation. Using Action Chunking with Transformers (ACT) as the base architecture, we train single-task models on bilateral wrench demonstrations and evaluate them on five contact-rich manipulation tasks. The wrench policy matches or outperforms position-based baselines across all tasks, with gains varying according to the degree of deliberate force regulation each task requires. Cross-condition ablations show that the bilateral data collection interface and the wrench action space each contribute independently to performance. To support further research, we will release over 1000 wrench-action demonstrations spanning these tasks on a companion website upon publication.