This website requires JavaScript.

S3E: A Large-scale Multimodal Dataset for Collaborative SLAM

Dapeng FengYuhua QiShipeng Zhong ...+4 Hongbo Chen
Oct 2022
摘要
With the advanced request to employ a team of robots to perform a taskcollaboratively, the research community has become increasingly interested incollaborative simultaneous localization and mapping. Unfortunately, existingdatasets are limited in the scale and variation of the collaborativetrajectories they capture, even though generalization betweeninter-trajectories among different agents is crucial to the overall viabilityof collaborative tasks. To help align the research community's contributionswith real-world multiagent ordinated SLAM problems, we introduce S3E, a novellarge-scale multimodal dataset captured by a fleet of unmanned ground vehiclesalong four designed collaborative trajectory paradigms. S3E consists of 7outdoor and 5 indoor scenes that each exceed 200 seconds, consisting of wellsynchronized and calibrated high-quality stereo camera, LiDAR, andhigh-frequency IMU data. Crucially, our effort exceeds previous attemptsregarding dataset size, scene variability, and complexity. It has 4x as muchaverage recording time as the pioneering EuRoC dataset. We also provide carefuldataset analysis as well as baselines for collaborative SLAM and singlecounterparts. Find data, code, and more up-to-date information athttps://github.com/PengYu-Team/S3E.
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?
0
被引用
笔记
问答