感知增强的机械手遥操作关键技术

发布时间：2019-02-11 12:03

【摘要】：随着机器人模仿技术的发展,机器人已经可以根据人类动作数据以实时或离线的方式模仿人类行为,采集人类动作的各种方法和设备也越来越成熟。机器人仿真技术的发展使得机器人自由度变得越来越多,机器人变得更加灵活,人类一些复杂的动作机器人可以很好地模仿。但是现有的研究缺少对于机器人模仿过程中机器人感知的反馈。本文针对缺少反馈的问题设计了一个感知增强的机械手遥操作系统,能够通过遥操作的方式远程控制机械手模仿操作者的手部动作,并采集机械手运行过程中的视觉和触觉信息反馈给人,增强机械手感知。本文的主要贡献:1)实现了一种基于Kinect和颜色标记的人类动作识别方式,可以获取到人手运动过程中的手部动作数据;2)基于采集的手部动作数据,设计并实现了一个手部动作数据到机械手指令的映射算法,并对映射效果进行了评估;3)基于Kinect和压力传感器,构建了机械臂运动过程中的反馈系统,包括视觉反馈和触觉反馈;4)基于感知增强的机械手遥操作系统布置了遥操作机械手移动物体的应用,并通过实验对系统进行了评估。
[Abstract]:With the development of robot imitation technology, robot can imitate human behavior in real time or offline according to human action data, and various methods and equipments for collecting human action are becoming more and more mature. With the development of robot simulation technology, the degree of freedom of robot becomes more and more, and the robot becomes more flexible. However, the existing research lacks feedback on robot perception in the process of robot imitation. In order to solve the problem of lack of feedback, this paper designs a sensing enhanced teleoperation system for manipulators, which can remotely control the manipulators' hand movements by remote operation. The visual and tactile information of the manipulator is collected to enhance the manipulator's perception. The main contributions of this paper are as follows: 1) A human motion recognition method based on Kinect and color tag is implemented, which can obtain the hand movement data in the process of hand movement; 2) based on the collected hand action data, a mapping algorithm from hand motion data to manipulator instruction is designed and implemented, and the mapping effect is evaluated. 3) based on Kinect and pressure sensor, a feedback system is constructed, including visual feedback and tactile feedback. 4) the teleoperation system of manipulator based on perceptual enhancement is applied to the mobile object of the manipulator, and the system is evaluated by experiments.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP241

【相似文献】