Estimating Motion Codes from Demonstration Videos
Estimating Motion Codes from Demonstration Videos
复制标题
DOI：
10.1109/iros45743.2020.9341065
复制
发表时间：
2020-07
期刊：
影响因子：
--
通讯作者：
Maxat Alibayev;D. Paulius;Yu Sun
中科院分区：
文献类型：
--
作者：
Maxat Alibayev;D. Paulius;Yu Sun
运动分类法可以将操作编码为二进制编码表示，我们将其称为运动代码。这些运动代码本质上代表了嵌入式空间中的操纵动作，描述了运动的机械特征，包括接触和轨迹类型。使用运动代码进行嵌入的主要优点是可以使用机器人相关特征更适当地定义运动，并且可以使用这些运动特征更合理地测量它们的距离。在本文中，我们开发了一个深度学习管道，以无监督的方式从演示视频中提取运动代码，以便这些视频中的知识可以正确表示并用于机器人。我们的评估表明，可以从 EPIC-KITCHENS 数据集中的动作演示中提取运动代码。
A motion taxonomy can encode manipulations as a binary-encoded representation, which we refer to as motion codes. These motion codes innately represent a manipulation action in an embedded space that describes the motion’s mechanical features, including contact and trajectory type. The key advantage of using motion codes for embedding is that motions can be more appropriately defined with robotic-relevant features, and their distances can be more reasonably measured using these motion features. In this paper, we develop a deep learning pipeline to extract motion codes from demonstration videos in an unsupervised manner so that knowledge from these videos can be properly represented and used for robots. Our evaluations show that motion codes can be extracted from demonstrations of action in the EPIC-KITCHENS dataset.