Active Next-Best-View Optimization for Risk-Averse Path Planning

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Khass, Amirhossein Mollaei, Liu, Guangyi, Pandey, Vivek, Jiang, Wen, Lei, Boshu, Daniilidis, Kostas, Motee, Nader
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866918156326404096
author Khass, Amirhossein Mollaei
Liu, Guangyi
Pandey, Vivek
Jiang, Wen
Lei, Boshu
Daniilidis, Kostas
Motee, Nader
author_facet Khass, Amirhossein Mollaei
Liu, Guangyi
Pandey, Vivek
Jiang, Wen
Lei, Boshu
Daniilidis, Kostas
Motee, Nader
contents Safe navigation in uncertain environments requires planning methods that integrate risk aversion with active perception. In this work, we present a unified framework that refines a coarse reference path by constructing tail-sensitive risk maps from Average Value-at-Risk statistics on an online-updated 3D Gaussian-splat Radiance Field. These maps enable the generation of locally safe and feasible trajectories. In parallel, we formulate Next-Best-View (NBV) selection as an optimization problem on the SE(3) pose manifold, where Riemannian gradient descent maximizes an expected information gain objective to reduce uncertainty most critical for imminent motion. Our approach advances the state-of-the-art by coupling risk-averse path refinement with NBV planning, while introducing scalable gradient decompositions that support efficient online updates in complex environments. We demonstrate the effectiveness of the proposed framework through extensive computational studies.
format Preprint
id arxiv_https___arxiv_org_abs_2510_06481
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Active Next-Best-View Optimization for Risk-Averse Path Planning
Khass, Amirhossein Mollaei
Liu, Guangyi
Pandey, Vivek
Jiang, Wen
Lei, Boshu
Daniilidis, Kostas
Motee, Nader
Robotics
Computer Vision and Pattern Recognition
Safe navigation in uncertain environments requires planning methods that integrate risk aversion with active perception. In this work, we present a unified framework that refines a coarse reference path by constructing tail-sensitive risk maps from Average Value-at-Risk statistics on an online-updated 3D Gaussian-splat Radiance Field. These maps enable the generation of locally safe and feasible trajectories. In parallel, we formulate Next-Best-View (NBV) selection as an optimization problem on the SE(3) pose manifold, where Riemannian gradient descent maximizes an expected information gain objective to reduce uncertainty most critical for imminent motion. Our approach advances the state-of-the-art by coupling risk-averse path refinement with NBV planning, while introducing scalable gradient decompositions that support efficient online updates in complex environments. We demonstrate the effectiveness of the proposed framework through extensive computational studies.
title Active Next-Best-View Optimization for Risk-Averse Path Planning
topic Robotics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2510.06481