首页> 外文会议>Proceedings of the IASTED technology conferences 2010 >TRAJECTORY LEARNING BASED ON CONDITIONAL RANDOM FIELDS FOR ROBOT PROGRAMMING BY DEMONSTRATION
【24h】

TRAJECTORY LEARNING BASED ON CONDITIONAL RANDOM FIELDS FOR ROBOT PROGRAMMING BY DEMONSTRATION

机译:基于条件随机域的机器人演示演示轨迹学习

获取原文
获取原文并翻译 | 示例

摘要

This work presents an approach for implementation ofrnconditional random fields (CRF) in transferring motorrnskills to robots. As a discriminative probabilistic model,rnCRF models directly the conditional probabilityrndistribution over label sequences for given observationrnsequences. Hereby, CRF was employed for segmentationrnand labeling of a set of demonstrated trajectories observedrnby a tracking sensor. The key points obtained by CRFrnsegmentation of the demonstrations were used forrngenerating a generalized trajectory for the taskrnreproduction. The approach was evaluated by simulationsrnof two industrial manufacturing applications.
机译:这项工作提出了一种在将运动技能传递给机器人时实现有条件随机场(CRF)的方法。作为判别概率模型,CRF直接针对给定的观察序列对标记序列上的条件概率分布进行建模。因此,CRF被用于分割和标记由跟踪传感器观察到的一组演示轨迹。通过演示的CRF分段得到的关键点用于生成任务再现的广义轨迹。该方法已通过Simulationsrnof在两个工业制造应用程序中进行了评估。

著录项

  • 来源
  • 会议地点 Cambridge MA(US);Cambridge MA(US);Cambridge MA(US);Cambridge MA(US);Cambridge MA(US)
  • 作者单位

    Department of Mechanical and Industrial Engineering, Ryerson University 350 Victoria Street, Toronto, Ont., M5B 2K3, Canada National Research Council Canada (NRC) - Institute for Aerospace Research (IAR) 5145 Decelles Avenue, Montreal, Que., H3T 2B2, Canadaa leksandar.vakanski@ryerson.ca;

    Department of Mechanical and Industrial Engineering, Ryerson University 350 Victoria Street, Toronto, Ont., M5B 2K3, Canada fsharifi@ryerson.ca;

    National Research Council Canada (NRC) - Institute for Aerospace Research (IAR) 5145 Decelles Avenue, Montreal, Que., H3T 2B2, Canada iraj.mantegh@cnrc-nrc.gc.ca;

    National Research Council Canada (NRC) - Institute for Aerospace Research (IAR) 5145 Decelles Avenue, Montreal, Que., H3T 2B2, Canada Andrew.Irish@cnrc-nrc.gc.ca;

  • 会议组织
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类 通信;计算技术、计算机技术;
  • 关键词

    robotics; programming by demonstration; industrial automation; artificial intelligence;

    机译:机器人技术;通过演示编程;工业自动化;人工智能;

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