Sampling-free obstacle gradients and reactive planning in Neural Radiance Fields


METADATA ONLY
Loading...

Date

2022-05-27

Publication Type

Conference Paper

ETH Bibliography

yes

Citations

Altmetric
METADATA ONLY

Data

Rights / License

Abstract

This work investigates the use of Neural implicit representations, specifically Neural Radiance Fields (NeRF), for geometrical queries and motion planning. We show that by adding the capacity to infer occupancy in a radius to a pre trained NeRF we are effectively learning an approximation to a Euclidean Signed Distance Field (ESDF). Even more, using backward differentiation of the network, we readily obtain the obstacle gradients that are integrated into policies for a Riemannian Motion Policies (RMP) framework. Thus, our findings allow for a sampling-free obstacle avoidance planning method in the implicit representation.

Publication status

published

Editor

Book title

Journal / series

Volume

Pages / Article No.

Publisher

Stanford University

Event

Workshop on "Motion Planning with Implicit Neural Representations of Geometry" at ICRA 2022

Edition / version

Methods

Software

Geographic location

Date collected

Date created

Subject

Organisational unit

03737 - Siegwart, Roland Y. / Siegwart, Roland Y.
02284 - NFS Digitale Fabrikation / NCCR Digital Fabrication

Notes

Funding

Related publications and datasets