Policy warps single demonstration to generate dexterous grasp motions
2nd Place Solution to the HANDS 2026 Workshop Challenge-Dexterous Grasp Motion Track: Single-Shot Trajectory Warping for Grasp Motion Generation
RoboticsArtificial IntelligenceComputer Vision and Pattern Recognition
Summary
Grasping objects with a complex robotic hand is tricky because it requires many precise movements in sequence. The authors found a way to teach a robot hand to adjust a single example of a good grasp to fit new objects and starting positions, instead of deciding every step from scratch. This makes the robot’s movements more reliable and easier to compute, even when handling objects it hasn’t seen before. Their method worked well in a challenge, achieving high success rates with both simple and tough tasks.
What this means in practice
- •For robotics engineers: Generate effective grasp motions for dexterous robotic hands using a warped demonstration to simplify control.
- •For robotics simulation developers: Create simulators that use single-shot motion warping policies to test grasping strategies on diverse virtual objects.
Authors
Muneeb A. Khan, Woojin Kim, Shinwoo Kim, Muhammad Munsif, Binod Bhattarai, Seungryul Baek
Abstract
This report describes our 2nd place solution to the HANDS 2026 workshop challenge (Dexterous Grasp Motion track) in conjunction with ECCV 2026. In this challenge, we address grasp motion generation for the 12-DoF LinkerHand O6, aiming to produce physically plausible reach-and-lift trajectories for unseen objects from randomized initial hand poses in simulation. This task is particularly challenging because each grasp requires a per-step policy to make approximately $70$ twelve-dimensional decisions, with errors accumulating over time, while test objects and physical dynamics may differ from those encountered during training. To address these challenges, we propose editing a single successful GraspM3 demonstration instead of generating the motion step by step: a policy observes the object once and outputs a 12-D warp of the demonstration, which is then replayed open-loop. Moreover, we train the warp policy with one-step PPO over all $4{,}824$ training objects in parallel. As a result, our method achieved success rates of $94.61\%$ on the easy track, the highest of all submissions, and $57.18\%$ on the hard track of the private test set.