.png){kind=icon}
.png){kind=icon}
Return
dGrasp: NeRF-Informed implicit grasp policies with supervised optimization slopes
.png){kind=icon}
OA
AI
DOI：10.1016/j.robot.2025.104921.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
We present dGrasp, an implicit grasp policy with an enhanced optimization landscape. This landscape is defined by a NeRF-informed grasp value function. The neural network representing this function is trained on simulated grasp demonstrations. During training, we use an auxiliary loss to guide not only the weight updates of this network but also the slope of the optimization landscape. This loss is computed on the demonstrated grasp trajectory and the gradients of the landscape. It requires second order optimization during training to incorporate valuable information from the trajectory and leads to facilitating the optimization process of the implicit policy. Experiments demonstrate that employing this auxiliary loss improves policies' performance in simulation as well as their zero-shot transfer to the real-world.
Keywords:
Implicit policy
Grasping
Implicit representation
Sim-to-real
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
5.2
Papers:
636
Citations:
1.0W