mindmap: Spatial Memory in Deep Feature Maps for 3D Action Policies

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Steiner, Remo, Millane, Alexander, Tingdahl, David, Volk, Clemens, Ramasamy, Vikram, Yao, Xinjie, Du, Peter, Pouya, Soha, Sheng, Shiwei
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866911195972239360
author Steiner, Remo
Millane, Alexander
Tingdahl, David
Volk, Clemens
Ramasamy, Vikram
Yao, Xinjie
Du, Peter
Pouya, Soha
Sheng, Shiwei
author_facet Steiner, Remo
Millane, Alexander
Tingdahl, David
Volk, Clemens
Ramasamy, Vikram
Yao, Xinjie
Du, Peter
Pouya, Soha
Sheng, Shiwei
contents End-to-end learning of robot control policies, structured as neural networks, has emerged as a promising approach to robotic manipulation. To complete many common tasks, relevant objects are required to pass in and out of a robot's field of view. In these settings, spatial memory - the ability to remember the spatial composition of the scene - is an important competency. However, building such mechanisms into robot learning systems remains an open research problem. We introduce mindmap (Spatial Memory in Deep Feature Maps for 3D Action Policies), a 3D diffusion policy that generates robot trajectories based on a semantic 3D reconstruction of the environment. We show in simulation experiments that our approach is effective at solving tasks where state-of-the-art approaches without memory mechanisms struggle. We release our reconstruction system, training code, and evaluation tasks to spur research in this direction.
format Preprint
id arxiv_https___arxiv_org_abs_2509_20297
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle mindmap: Spatial Memory in Deep Feature Maps for 3D Action Policies
Steiner, Remo
Millane, Alexander
Tingdahl, David
Volk, Clemens
Ramasamy, Vikram
Yao, Xinjie
Du, Peter
Pouya, Soha
Sheng, Shiwei
Robotics
End-to-end learning of robot control policies, structured as neural networks, has emerged as a promising approach to robotic manipulation. To complete many common tasks, relevant objects are required to pass in and out of a robot's field of view. In these settings, spatial memory - the ability to remember the spatial composition of the scene - is an important competency. However, building such mechanisms into robot learning systems remains an open research problem. We introduce mindmap (Spatial Memory in Deep Feature Maps for 3D Action Policies), a 3D diffusion policy that generates robot trajectories based on a semantic 3D reconstruction of the environment. We show in simulation experiments that our approach is effective at solving tasks where state-of-the-art approaches without memory mechanisms struggle. We release our reconstruction system, training code, and evaluation tasks to spur research in this direction.
title mindmap: Spatial Memory in Deep Feature Maps for 3D Action Policies
topic Robotics
url https://arxiv.org/abs/2509.20297