aiqbench analysis
This research advances robotic manipulation by shifting policy outputs from positional targets to direct wrench predictions, enabling more precise force regulation in contact-rich tasks. By leveraging bilateral teleoperation data and the ACT architecture, the study demonstrates that pure force-domain learning can match or exceed position-based baselines across diverse scenarios. The release of over 1000 demonstration datasets will facilitate further exploration into how deliberate human force input shapes autonomous robot behavior.
Wrench-ACT: Enhancing Robot Policies for Contact Rich Behavior Using Direct Wrench Control
arXiv:2609.37552v1 Announce Type: new 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.
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.
Robotics (cs.RO)
arXiv:2609.37552 [cs.RO]
(or arXiv:2609.37552v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2609.37552 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 284 words
Related coverage
PRISM: Polynomial Representations for Interaction-Structured Motor Control
arXiv · Jul 28, 2026
Imp-ACT: Adaptive Impedance Control and Action Chunking with Transformers to Learn Contact-Rich Manipulation from Demonstrations
arXiv · Sep 28, 2026
NVIDIA GTC Taipei at COMPUTEX: Live Updates on What’s Next in AI
NVIDIA · May 21, 2026
TORL-VLA: Tactile Guided Online Reinforcement Learning for Contact-Rich Manipulation
arXiv · Jun 11, 2026