This website requires JavaScript.

Circular Accessible Depth: A Robust Traversability Representation for UGV Navigation

Shikuan XieRan SongYuenan ZhaoXueqin HuangYibin LiWei Zhang
Dec 2022
摘要
In this paper, we present the Circular Accessible Depth (CAD), a robusttraversability representation for an unmanned ground vehicle (UGV) to learntraversability in various scenarios containing irregular obstacles. To predictCAD, we propose a neural network, namely CADNet, with an attention-basedmulti-frame point cloud fusion module, Stability-Attention Module (SAM), toencode the spatial features from point clouds captured by LiDAR. CAD isdesigned based on the polar coordinate system and focuses on predicting theborder of traversable area. Since it encodes the spatial information of thesurrounding environment, which enables a semi-supervised learning for theCADNet, and thus desirably avoids annotating a large amount of data. Extensiveexperiments demonstrate that CAD outperforms baselines in terms of robustnessand precision. We also implement our method on a real UGV and show that itperforms well in real-world scenarios.
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