analiza aiqbench
This research addresses partial observability in robotic navigation by introducing a coupled learning scheme that jointly predicts Truncated Unsigned Distance Field (TUDF) representations and voxel-wise occupancy maps from partial LiDAR data. The bidirectional synergy between these geometric formulations enhances obstacle boundary reasoning, enabling seamless integration into trajectory planning without post-processing. This advancement represents a significant step forward in machine learning for robotics, specifically optimizing scene completion for downstream path-planning tasks.
Planning Oriented 3D Scene Completion via Coupled TUDF Occupancy Representation Learning from Partial Observations
arXiv:2609.36543v1 Announce Type: new Abstract: Partial observability remains a fundamental challenge in robotic navigation, where limited sensor coverage and occlusions leave large portions of the environment unobserved. Existing scene completion methods primarily focus on improving incomplete mapping or reconstructing partially observed 3D structures, but rarely investigate how scene completion can be designed to benefit downstream tasks such as path planning. In this work, we propose a path-planning-oriented 3D scene completion framework that moves beyond pure occupancy modeling toward a coupled geometric formulation. Specifically, given partial LiDAR observations as input, the proposed framework jointly predicts completed Truncated Unsigned Distance Field (TUDF)-based continuous geometric representations and voxel-wise occupancy maps. This coupled representation allows the network to better reason about obstacle boundaries and free-space geometry. To fully exploit the synergy between the two representations, we introduce a bidirectionally coupled learning scheme, where TUDF features provide dense geometric guidance to improve occupancy reconstruction, while occupancy features in turn offer complementary structural constraints that refine distance-field estimation. Consequently, the proposed network directly predicts complete occupancy and TUDF representations, allowing seamless integration of TUDF into trajectory planning without post-processing. Extensive experiments on unseen environments demonstrate that the proposed method consistently improves both geometric reconstruction quality and downstream planning performance.
Abstract:Partial observability remains a fundamental challenge in robotic navigation, where limited sensor coverage and occlusions leave large portions of the environment unobserved. Existing scene completion methods primarily focus on improving incomplete mapping or reconstructing partially observed 3D structures, but rarely investigate how scene completion can be designed to benefit downstream tasks such as path planning. In this work, we propose a path-planning-oriented 3D scene completion framework that moves beyond pure occupancy modeling toward a coupled geometric formulation. Specifically, given partial LiDAR observations as input, the proposed framework jointly predicts completed Truncated Unsigned Distance Field (TUDF)-based continuous geometric representations and voxel-wise occupancy maps. This coupled representation allows the network to better reason about obstacle boundaries and free-space geometry. To fully exploit the synergy between the two representations, we introduce a bidirectionally coupled learning scheme, where TUDF features provide dense geometric guidance to improve occupancy reconstruction, while occupancy features in turn offer complementary structural constraints that refine distance-field estimation. Consequently, the proposed network directly predicts complete occupancy and TUDF representations, allowing seamless integration of TUDF into trajectory planning without post-processing. Extensive experiments on unseen environments demonstrate that the proposed method consistently improves both geometric reconstruction quality and downstream planning performance.
Robotics (cs.RO)
arXiv:2609.36543 [cs.RO]
(or arXiv:2609.36543v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2609.36543 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 291 cuvinte
Subiecte conexe
Auto-JEPA: A Latent World Model of Continuous Intent for End-to-End Autonomous Driving
arXiv · 3 aug. 2026
OccWAM: Occupancy World-Action Models with Joint Scene-to-Ego Flow Matching
arXiv · 14 oct. 2026
Accelerating BEV Pooling on NVIDIA GPUs for Physical AI Applications
NVIDIA · 24 iun. 2026
Developing Healthcare Robotics with GPU-Native Medical Physics Simulation
NVIDIA · 28 iul. 2026