摘要
为快速确定钣金件抓取位置,提高抓取自动化程度,提出一种基于概率热力图的钣金件抓取点生成方法。基于钣金件与端拾器投影计算单因素概率热力图,评估钣金件与端拾器重叠程度、钣金件几何中心位置、钣金件表面起伏程度;融合各单因素概率热力图对概率热力图进行随机网格划分,并以数值最高处作为抓取位置;最后通过有限元仿真验证抓取点有效性。结果表明,所提抓取点生成与评估方法可以适应不同抓取工况,单次概率热力图计算及抓取点生成需1.58 s。Q345b钣金件被抓取最大形变为7.993 mm,最大等效应力为158.42 MPa,小于钣金件材料屈服强度。
To determine the grasping points of sheet metal parts quickly and improve the degree of automation of grasping,a probability heat map-based grasping point generation and evaluation method for sheet metal parts was proposed.The single-factor probability heat maps were calculated based on the projection of sheet metal parts and end pickers to evaluate the degree of overlapping between the sheet metal parts and the end pickers,the geometric center of sheet metal parts and the surface smoothness of sheet metal parts.Then,single-factor probability heat maps were fused.The fused probability heat map was randomly meshed and the locations of highest values were selected as grasping points.Finally,finite element analysis was utilized to verify the validity of the generated grasping points.The results showed that the proposed grasping point generation and evaluation method was applicable to different grasping working conditions.The runtime of probability heat map calculation and grasping point generation was 1.58s.The maximum total deformation of the gripped Q345b sheet metal part was 7.993mm,and the maximum equivalent stress was 158.42MPa,which was smaller than the yield strength of the material of the sheet metal part.
作者
黄洪鑫
王欣
陈立
张执南
HUANG Hongxin;WANG Xin;CHEN Li;ZHANG Zhinan(School of Mechanical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China;Tingbai(Yangzhou)Intelligent Technology Co.,Ltd.,Yangzhou 225200,China)
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2024年第7期2306-2318,共13页
Computer Integrated Manufacturing Systems
基金
上海市科技计划资助项目(23PJ1405700)。
关键词
钣金件
概率热力图
抓取点生成
有限元分析
sheet metal part
probability heat map
grasping point generation
finite element analysis
作者简介
黄洪鑫(1999-),男,湖南益阳人,硕士研究生,研究方向:智能制造、机器视觉及其工业应用,E-mail:huanghongxin@sjtu.edu.cn;王欣(1994-),女,河北邯郸人,助理研究员,博士,研究方向:数模驱动的智能设计、供应链管理,E-mail:wangxinmandy@sjtu.edu.cn;陈立(1982-),男,江苏扬州人,硕士,研究方向:金属加工设备研发、智能机器人研发,E-mail:13472826733@126.com;通讯作者:张执南(1978-),男,辽宁盘锦人,教授,博士,博士生导师,研究方向:数据驱动的设计、摩擦信息学,E-mail:zhinanz@sjtu.edu.cn。
