首页 | 本学科首页   官方微博 | 高级检索  
     

机器人接触状态下作业的技能获取及融合
引用本文:王琴,杨汝清,张伟军. 机器人接触状态下作业的技能获取及融合[J]. 上海交通大学学报, 2004, 38(1): 47-51
作者姓名:王琴  杨汝清  张伟军
作者单位:上海交通大学,机器人研究所,上海,200030;上海交通大学,机器人研究所,上海,200030;上海交通大学,机器人研究所,上海,200030
基金项目:国家自然科学基金项目(69975044),国家高技术研究发展计划(863)项目(863-512-9820-04)
摘    要:为使机器人获得人的基于力信息的作业技能,提出了用离散接触状态的变迁来建立作业的技能模型.用元素接触形式(ECF)描述定性接触状态,根据不同ECF的力信息聚类,采用改进的FCM聚类算法对示教所得的力信息进行分类,从而从人的示教中自动获取作业技能.为了对多次示教的作业技能进行融合,用Petri网进行建模,并提出根据技能模型描述状态变迁的矩阵方法.技能最终以面向对象的语言Visual C 描述.

关 键 词:模糊C均值聚类  Petri  机器人  接触状态  元素接触形式
文章编号:1006-2467(2004)01-0047-05
修稿时间:2003-01-03

Robot Performance Skill Learning and Merging for Tasks in Contact States
WANG Qin,YANG Ru-qing,ZHANG Wei-jun. Robot Performance Skill Learning and Merging for Tasks in Contact States[J]. Journal of Shanghai Jiaotong University, 2004, 38(1): 47-51
Authors:WANG Qin  YANG Ru-qing  ZHANG Wei-jun
Abstract:The motivation behind this work is to transfer force-based skill to robots. Firstly the skill was modeled as a sequence of discrete contact states and transformation of contact states. Contact state was described with element contact formation(ECF). For different ECF has different force clustering, a modified FCM was used to learn the skill automatically from human demonstration. Different skills were merged with Petri net. Matrix based method was proposed to realize the transfer of contact states. Finally the skill model was implemented with object-orient language.
Keywords:fuzzy C mean(FCM)  Petri  robot  contact state  element contact formation(ECF)
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号