PulseAugur
EN
LIVE 12:16:57
中文(ZH) ICRA 2026| 论文评述Kilometer-Scale GNSS-Denied UAV Navigation

Drone navigation system wins challenge using LiDAR and prior maps

A Czech team developed a novel system for long-range drone navigation in GPS-denied environments, winning the SPRIN-D Funke challenge. The system uses a clustered particle filter to fuse LiDAR-generated heightmaps with prior geographical data, correcting for odometry drift without relying on GNSS. It successfully navigated kilometers through varied terrain on a CPU-only hardware setup, demonstrating significant improvements over standard odometry. AI

IMPACT Demonstrates advanced AI techniques for robust drone navigation in challenging, real-world conditions, potentially impacting autonomous systems in logistics and surveillance.

RANK_REASON The article details a specific technical achievement in drone navigation for a competition, focusing on the methodology and results. [lever_c_demoted from research: ic=1 ai=1.0]

Read on 雷峰网 (Leiphone) →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Drone navigation system wins challenge using LiDAR and prior maps

COVERAGE [1]

  1. 雷峰网 (Leiphone) TIER_1 中文(ZH) ·

    ICRA 2026 | Paper Review: Kilometer-Scale GNSS-Denied UAV Navigation

    <p><br /></p><p>原文作者:Michal Werner, David Čapek 等4名</p><p>原文链接:https://www.themoonlight.io</p><p>雷峰网注:该系统由捷克理工大学(CTU in Prague)著名的多机器人系统小组(MRS)开发。它是针对德国&nbsp;SPRIN-D Funke 挑战赛(完全自主飞行挑战赛)&nbsp;专门研发的夺冠方案。比赛要求无人机在没有 GNSS 信号、没有预先构建的稠密地图的情况下,在 25 米以下(AGL)的低空自主飞行 9 公里进行航点导航。</p><p><br …