遥操作
计算机科学
计算机视觉
人工智能
对象(语法)
机器人
任务(项目管理)
脑电图
接口(物质)
触觉技术
人机交互
块(置换群论)
模拟
心理学
工程类
几何学
数学
系统工程
气泡
精神科
最大气泡压力法
并行计算
作者
Minki Kim,Myoung‐Su Choi,Ga-Ram Jang,Ji‐Hun Bae,Hyung‐Soon Park
标识
DOI:10.3389/fnbot.2023.1293878
摘要
This paper presents a teleoperation system of robot grasping for undefined objects based on a real-time EEG (Electroencephalography) measurement and shared autonomy. When grasping an undefined object in an unstructured environment, real-time human decision is necessary since fully autonomous grasping may not handle uncertain situations. The proposed system allows involvement of a wide range of human decisions throughout the entire grasping procedure, including 3D movement of the gripper, selecting proper grasping posture, and adjusting the amount of grip force. These multiple decision-making procedures of the human operator have been implemented with six flickering blocks for steady-state visually evoked potentials (SSVEP) by dividing the grasping task into predefined substeps. Each substep consists of approaching the object, selecting posture and grip force, grasping, transporting to the desired position, and releasing. The graphical user interface (GUI) displays the current substep and simple symbols beside each flickering block for quick understanding. The tele-grasping of various objects by using real-time human decisions of selecting among four possible postures and three levels of grip force has been demonstrated. This system can be adapted to other sequential EEG-controlled teleoperation tasks that require complex human decisions.
科研通智能强力驱动
Strongly Powered by AbleSci AI