A Motion Taxonomy for Manipulation Embedding
A Motion Taxonomy for Manipulation Embedding
复制标题
DOI：
10.15607/rss.2020.xvi.045
复制
发表时间：
2020-07
期刊：
影响因子：
--
通讯作者：
D. Paulius;Nicholas Eales;Yu Sun
中科院分区：
文献类型：
--
作者：
D. Paulius;Nicholas Eales;Yu Sun
为了从机械的角度表示运动，本文使用运动分类法来探索运动嵌入。通过这种分类，可以将操作描述和表示为称为运动代码的二进制字符串。运动代码捕获力学属性，例如接触类型和轨迹，这些属性应该用于定义用于深度学习和强化学习的适当的运动之间的距离度量或损失函数。运动代码还可以用于合并别名或聚集共享相似特性的运动类型。以现有的数据集为参照，我们讨论了如何根据直觉和真实数据创建运动代码并将其分配给日常生活活动中常见的动作。运动代码与来自预训练的word2vec模型的矢量进行了比较，我们表明运动代码保持的距离与操作的现实情况非常接近。
To represent motions from a mechanical point of view, this paper explores motion embedding using the motion taxonomy. With this taxonomy, manipulations can be described and represented as binary strings called motion codes. Motion codes capture mechanical properties, such as contact type and trajectory, that should be used to define suitable distance metrics between motions or loss functions for deep learning and reinforcement learning. Motion codes can also be used to consolidate aliases or cluster motion types that share similar properties. Using existing data sets as a reference, we discuss how motion codes can be created and assigned to actions that are commonly seen in activities of daily living based on intuition as well as real data. Motion codes are compared to vectors from pre-trained Word2Vec models, and we show that motion codes maintain distances that closely match the reality of manipulation.