引用本文:苏虎,徐德,陈国栋,方灶军.基于两阶段多目标模糊决策的乒乓球机器人击球点选择[J].控制理论与应用,2013,30(5):597~603.[点击复制]
SU Hu,XU De,CHEN Guo-dong,FANG Zao-jun.Striking position selection based on two-step multi-purpose fuzzy decision method for robotic table tennis[J].Control Theory and Technology,2013,30(5):597~603.[点击复制]
基于两阶段多目标模糊决策的乒乓球机器人击球点选择
Striking position selection based on two-step multi-purpose fuzzy decision method for robotic table tennis
摘要点击 2403  全文点击 2426  投稿时间：2012-08-27  修订日期：2012-12-14
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2013.20907
  2013,30(5):597-603
中文关键词  击球点选择  模糊决策  乒乓球机器人  期望出球速度  球拍参数
英文关键词  striking position selection  fuzzy decision  table tennis robot  desired outgoing velocity  racket parameter
基金项目  国家自然科学基金资助项目(61273337, 61075035).
作者单位E-mail
苏虎* 中国科学院 自动化研究所 复杂系统管理与控制国家重点实验室 suhumathe@yahoo.com.cn 
徐德 中国科学院 自动化研究所 精密感知与控制研究中心  
陈国栋 中国科学院 自动化研究所 复杂系统管理与控制国家重点实验室  
方灶军 中国科学院 自动化研究所 复杂系统管理与控制国家重点实验室  
中文摘要
      对乒乓球机器人而言, 回球时根据来球状态选择合适的击球点, 可以降低回球难度, 提高回球成功率. 本文提出了一种利用多目标模糊决策确定击球点的方法. 首先, 根据机器人最优运动区间和来球速度确定搜索区域. 然后, 定义搜索区域内的相关轨迹点为决策论域. 在此决策论域内, 根据回球目标参数分别计算期望出球速度, 并根据球拍参数定义了3个能量函数, 分别代表乒乓球被击打前后的能量和动量变化. 以此能量函数为指标函数决策得到优化的击球点. 最后, 在不同的实验情况下, 对比了本文算法与固定高度平面算法, 验证了本文算法的有效性.
英文摘要
      For robotic table tennis, selecting the optimal striking position in the returning process according to the ball’s trajectory, not only reduces the difficulties in returning the coming ball, but also increases the returnable success rate. A two-step multi-purpose fuzzy decision making method is proposed for striking-position selection. Firstly, according to robot’s optimum range of motion and the coming ball’s velocity, the searching region containing the striking position is determined. Then, the positions of the ball in the searching region are regarded as decision-making universe. In the universe, the desired outgoing velocity is respectively calculated with the target returning parameters. Three energy functions are defined based on the parameters of racket, representing energetic and momentum changes of the ping-pong ball in the striking process. The functions are used as index functions for striking point selection in the searching region. Finally, comparison experiments of the proposed method with conventional ones are well conducted under different experimental conditions, demonstrating effectiveness of the proposed method.