Research Deformable Linear Objects
AssistDLO: Assistive Teleoperation for Deformable Linear Object Manipulation
arXiv preprint arXiv:2605.06323, 2026
@misc{guler2026assistdlo,
title = {{AssistDLO}: Assistive Teleoperation for Deformable Linear Object Manipulation},
author = {Guler, Berk and Manschitz, Simon and Pompetzki, Kay and Peters, Jan},
year = {2026},
eprint = {2605.06323},
archivePrefix = {arXiv}
}Manipulating Deformable Linear Objects (DLOs) is challenging in robotics due to their infinite-dimensional configuration space and complex nonlinear dynamics. In teleoperation, depth uncertainty hinders state perception and reaction. AssistDLO addresses this challenge as an assistive teleoperation framework for DLO manipulation that combines real-time multi-view state estimation, visual assistance (VA), and a geometry-aware shared-autonomy controller based on Control Barrier Functions (SA-CBF). While traditional shared autonomy methods often rely on simple geometric attractors and may fail to preserve DLO geometry, SA-CBF acts as a geometry-aware funnel, facilitating precise grasping while preserving the operator's high-level authority. The framework is evaluated in a bimanual knot-untangling user study (N = 22) using ropes with varying length and rigidity. Results show that the effectiveness of the assistance depends strongly on operator expertise and DLO properties. SA-CBF provides the strongest gains for naive users, acting as a skill equalizer that increases task success from 71% to 88%, and is effective for stiffer ropes. Conversely, expert users prefer VA, and highly compliant, long ropes benefit more from visual support than localized action assistance. Ultimately, these findings demonstrate that effective DLO teleoperation cannot rely on a fixed strategy, highlighting the critical need for adaptive, user-aware, and material-aware shared autonomy.
AI-generated summary · may contain mistakes, see the paper
Ropes and cables are hard to teleoperate: their shape is high-dimensional, and the operator has to judge depth from a 2D video stream. AssistDLO keeps the human in charge of task-level decisions and helps with perception and precise grasping:
- Rope state estimation. Two wrist-mounted RGB-D cameras trace the rope (SAM 2 segmentation and a Voronoi-based centre line), fused into a 3D point sequence in real time.
- Visual assistance (VA). The estimated rope is shown as a registered 3D overlay in the operator's VR headset, adding stereo depth cues.
- Action assistance (SA-CBF). A control barrier function defines a height-field barrier above the table with approach funnels around rope points, so the gripper can descend only through these funnels and grasp precisely without sweeping through other parts of the rope.
In a 22-participant knot-untangling study, SA-CBF helped novices most (71% → 88% success) and worked best on stiffer ropes, while experts and long, compliant ropes benefited more from visual assistance, so the right assistance depends on the user and the material.
Try it: untangle the knot
Untie a loose overhand knot by grasping and moving strands, as the participants did with a bimanual robot. Click a strand to grasp it, drag to move it and release to put it down; you win when the rope no longer crosses itself. Switch on SA-CBF to get the grasp assistance from the paper. Stuck? The knot comes undone when a free end gets back out from under the strand that holds it, and the Hint button shows where.
A browser game inspired by the study task. The rope is a small simulation with gravity, table friction and self-collision, so strands cannot pass through each other. As in the study, the knot counts as untied when the rope no longer crosses itself. SA-CBF is shown in simplified form: a barrier keeps the gripper above the rope except in funnels (blue) around rope points, so a click near the rope is guided onto the nearest point, the upper strand at a crossing, and a click away from the rope does not make the gripper descend.
Key results
- 22participants, 352 knot-untangling trials across 4 ropes
- 71% → 88%task success for novice operators with SA-CBF
- 17–21%shorter completion time on successful trials with any assistance mode
My contributions
- Designed the
AssistDLOframework, combining real-time multi-view DLO state estimation with a geometry-aware assistance controller. - Formulated the
SA-CBFcontroller as a geometry-aware safety funnel around the estimated rope for precise, collision-aware grasping. - Designed and ran the bimanual knot-untangling user study (N = 22) across ropes of varying length and rigidity, comparing visual and action assistance against unassisted teleoperation.
- Analyzed how the effectiveness of assistance depends on operator expertise and rope properties.