Abstract
Swarm UAV autonomous flight for Long-Horizon (LH) tasks is crucial for advancing the low-altitude economy. However, the maximum trajectory length in existing swarm UAV autonomous flight datasets is limited to 15 seconds per flight path, which fails to explore LH tasks. This paper presents U2UData-2, the first large-scale swarm UAV autonomous flight dataset for LH tasks and the first scalable data collection platform. The dataset is captured by 15 UAVs in autonomous collaborative flights for LH tasks, comprising 12 scenes (weather and terrain combination), 720 traces, 120 hours (each trace 10 minutes), 4.32M LiDAR frames, and 12.96M RGB frames. They also include brightness, temperature, humidity, smoke, and airflow values covering all flight routes. The data collection platform supports the customization of simulators, UAVs, sensors, flight algorithms, formation modes, and LH tasks. Through its visual control window, U2UData-2 allows users to collect customized datasets through one-click deployment online and to verify algorithms by closed-loop simulation. U2UData-2 can be found at https://fengtt42.github.io/U2UData-2/.
Real-world Mapping Simulator
Scalable Simulator
U2UData-2 Dataset Setting
U2UData-2 Dataset Analysis
Long-Horizon Tasks
Cite
U2UData is nominated for the Best Paper in ACM Multimedia 2024.
author = {Feng, Tongtong and Wang, Xin and Han, Feilin and Zhang, Leping and Zhu, Wenwu},
title = {U2UData: A Large-scale Cooperative Perception Dataset for Swarm UAVs Autonomous Flight},
year = {2024},
booktitle = {Proceedings of the 32nd ACM International Conference on Multimedia},
pages = {7600–7608},
numpages = {9}
}
Team
