摘要
针对手指逆运动学问题,采用了基于贝叶斯网理论的仿生学研究方法.首先对人类手指进行了运动学建模,并通过多组实验,采样记录人手在抓取时,食指各关节在不同时刻的角度;其次,以采样数据为样本,进行贝叶斯网络学习,获得各关节角度之间的内在关系模型;最后将该模型应用于气动柔性手指控制系统.实验表明,该模型解决了多自由度机器人逆运动学冗余度的同时,使得气动柔性手指的运动更接近人类.
A biomimetic approach based on Bayesian network was proposed for solving the redundancy problem of flexible pneumatic finger.First,the kinematics of finger was investigated and modeled.With multiple sets of experiments,the finger's joint angle was recorded during the procedure of grabbing.Then,with Bayesian network,the dependency model of joint angle was described.At last,the model was used in flexible pneumatic finger's control system.Through experiment,it is proved that the proposed model results in human-like configurations while solving inverse kinematics algorithms.
出处
《中国计量学院学报》
2012年第2期156-159,共4页
Journal of China Jiliang University
关键词
仿生学
贝叶斯网
逆运动学
冗余度
biomimetic
Bayesian network
reverse kinematics
redundant
