Learning High-DOF Reaching-and-Grasping via Dynamic Representation of Gripper-Object Interaction

She, Qijin; Hu, Ruizhen; Xu, Juzhan; Liu, Min; Xu, Kai; Huang, Hui

doi:10.48550/arxiv.2204.13998

Cited by 1 publication

(2 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Generating high-quality human grasps remains challenging due to the complex geometry and complicated skeletal constraints. Physics-based methods [7,23,28,33,39,40] were favored for in-hand manipulation synthesis since the generated motion was physically plausible. IBS [33] presented a novel representation of hand-object interaction and leveraged reinforcement learning (RL) methods with execution success, and geometric measure [11] rewards to generate successful grasping motion.…”

Section: Physics-based Object Manipulation Synthesismentioning

confidence: 99%

“…Physics-based methods [7,23,28,33,39,40] were favored for in-hand manipulation synthesis since the generated motion was physically plausible. IBS [33] presented a novel representation of hand-object interaction and leveraged reinforcement learning (RL) methods with execution success, and geometric measure [11] rewards to generate successful grasping motion. D-Grasp [7] developed its grasping policy based on the physical attributes of hand and object, including angles and velocities.…”

Section: Physics-based Object Manipulation Synthesismentioning

confidence: 99%

See 1 more Smart Citation

CAMS: CAnonicalized Manipulation Spaces for Category-Level Functional Hand-Object Manipulation Synthesis

Zheng¹,

Zheng²,

Fang³

et al. 2023

Preprint

View full text Add to dashboard Cite

Figure 1. Given a sequence of manipulation goals, our method can generate realistic and diverse functional manipulation motions consistent with the goal sequence. The motions are expressed in snapshots at several keyframes. On the left side, we show a goal sequence of opening a laptop and two different manipulation patterns that can be generated by our method. On the right side, we show our generation results of three manipulation tasks corresponding to different object categories.

show abstract