SeqGrasp: Sequential Multi-Object Dexterous Grasp Generation

Haofei Lu, Yifei Dong, Zehang Weng, Jens Lundell, Florian T. Pokorny, Danica Kragic

Keywords: robot grasping; multi-object grasping

TL;DR: We introduce the sequential multi-object robotic grasp sampling algorithm SeqGrasp that can robustly synthesize stable grasps on diverse objects using the robotic hand’s partial Degrees of Freedom (DoF).

Abstract:

We introduce the sequential multi-object robotic grasp sampling algorithm SeqGrasp that can robustly synthesize stable grasps on diverse objects using the robotic hand’s partial Degrees of Freedom (DoF). We experimentally evaluate SeqGrasp against the state-of-the-art non-sequential multi-object grasp generation method MultiGrasp in simulation and on a real robot. The experimental results demonstrate that
SeqGrasp reach an 8.71%-43.33% higher grasp success rate than MultiGrasp.

Submission Number: 3