• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>IEEE International Conference on Robotics and Biomimetics >Learning to control space robots with flexible appendages using model-based policy search
【24h】

Learning to control space robots with flexible appendages using model-based policy search

机译：学习使用基于模型的策略搜索来控制带有灵活附件的空间机器人

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Space robot is one of the promising approach to perform space tasks. Modeling and controlling of space robots is complicated caused by high dynamics coupling between the manipulator and the target spacecraft. Beyond that, vibration of flexible appendages mounted on the base of space robot can be excited, which is not expected for the precise control. After capturing the target satellite, the manipulator will be controlled to move from the capturing configuration to the berthing configuration, which is conveniently servicing the target. In this paper, we propose to learn the control policy of space robot directly from observed trajectories. A Gaussian regression model fitted to observed trajectories is used to approximate dynamics of space robotic system, based on which the control policy is learned. Plane space robot systems are used to verify the effectiveness of the presented approach. The simulation results indicate that the proposed method can well achieve the coupling control of the manipulator and the base. Moreover, the space robot can successfully berth the target while suppressing the vibrations of the structure.
机译：太空机器人是执行太空任务的有前途的方法之一。由于操纵器与目标航天器之间的高动力学耦合，导致空间机器人的建模和控制变得复杂。除此之外，还可以激发安装在太空机器人基座上的柔性附件的振动，这对于精确控制是无法预期的。在捕获目标卫星之后，将控制机械手从捕获配置移动到停泊配置，这很方便地为目标提供服务。在本文中，我们建议直接从观察到的轨迹中学习空间机器人的控制策略。拟合观测轨迹的高斯回归模型用于近似空间机器人系统的动力学，在此基础上学习控制策略。平面空间机器人系统用于验证所提出方法的有效性。仿真结果表明，该方法可以很好地实现机械手与基座的耦合控制。此外，太空机器人可以在抑制结构振动的同时成功停泊目标。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号