检索规则说明:AND代表“并且”;OR代表“或者”;NOT代表“不包含”;(注意必须大写,运算符两边需空一格)
检 索 范 例 :范例一: (K=图书馆学 OR K=情报学) AND A=范并思 范例二:J=计算机应用与软件 AND (U=C++ OR U=Basic) NOT M=Visual
作 者:王永志[1] 张禹[1] 赵蕾 吴利平[1] 张倩 WANG Yong-zhi;ZHANG Yu;ZHAO Lei;WU Li-ping;ZHANG Qian(School of Mechanical Engineering,Shenyang University of Technology,Shenyang 110870,China;College of Science,Shenyang University of Chemical Technology,Shenyang 110142,China)
机构地区:[1]沈阳工业大学机械工程学院,沈阳110870 [2]沈阳化工大学理学院,沈阳110142
出 处:《组合机床与自动化加工技术》2022年第8期152-155,159,共5页Modular Machine Tool & Automatic Manufacturing Technique
基 金:国家自然科学基金(51775354,52005344);辽宁省科技厅重点实验室联合基金(2020-KF-12-07)。
摘 要:为了提高生产线的柔性度与机器人的智能程度,提出了一种基于深度强化学习的多视角智能装配策略。该策略利用多视角视觉感知解决装配中因遮挡导致的特征缺失,并使用触觉感知信息作为视觉感知的预测标签;基于深度Q网络和深度确定性策略梯度算法设计了装配预测网络,并采用优先经验回放来减少训练中的非平稳分布与样本相关性。训练后的机器人装配效率提升了12.83%,实验结果表明该策略能够有效提高装配效率并改善装配效率的稳定性。To improve the flexibility of the production line and the intelligence level of the robot,an intelligent assembly strategy of multi-view based on deep reinforcement learning is proposed.This strategy uses multi-view visual perception to solve the missing features due to occlusion in assembly and utilizes tactile perception information as the predicted label of visual perception;the network of assembly prediction was designed based on the deep Q network and the deep deterministic strategy gradient algorithm and uses the priority experience replay to reduce the non-stationary distribution and sample correlation in training.The assembly efficiency of the trained robot is increased by 12.83%.The experimental results demonstrate that the proposed assembly strategy can effectively improve the assembly efficiency and the stability of the assembly efficiency.
分 类 号:TH165[机械工程—机械制造及自动化] TG502[金属学及工艺—金属切削加工及机床]
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在链接到云南高校图书馆文献保障联盟下载...
云南高校图书馆联盟文献共享服务平台 版权所有©
您的IP:18.119.122.86