A brief review of affordance in robotic manipulation research
A brief review of affordance in robotic manipulation research
复制标题
DOI：
10.1080/01691864.2017.1394912
复制
发表时间：
2017-10
影响因子：
2
通讯作者：
中科院分区：
文献类型：
--
作者：
摘要 本文简要回顾了机器人技术可供性研究，特别关注其在抓取和操纵物体方面的应用。可供性的概念可能是实现类人高级操纵智能的关键。首先，我们讨论可供性的概念，同时与机器人技术的应用相关联。然后，我们深入探索利用可供性进行机器人操作应用的研究，例如对象识别、抓取和对象操作（包括工具使用）。他们通过多种方式获取和使用可供性，例如向人类学习、使用模拟和现实世界的执行。此外，我们还展示了我们当前的工作，即用于高级操纵智能的云数据库。该数据库积累了与操作任务执行相关的各种数据，并将成为利用各种可供性技术的开放平台。图解摘要
Abstract This paper presents a brief review of affordance research in robotics, with special concentrations on its applications in grasping and manipulation of objects. The concept of affordance could be a key to realize human-like advanced manipulation intelligence. First, we discuss the concept of affordance while associating with the applications in robotics. Then, we intensively explore the studies that utilize affordance for robotic manipulation applications, such as object recognition, grasping, and object manipulation including tool-use. They obtain and use affordance by several ways like learning from human, using simulation, and real-world execution. Moreover, we show our current work, which is a cloud database for advanced manipulation intelligence. The database accumulates various data related to manipulation task execution and will be an open platform to leverage various affordance techniques. Graphical Abstract