智能拣选工作台中的人机协作研究  

Research on man-machine cooperation in intelligent picking workbench

作　　者：张贺龙 吴洪明[1] Zhang Helong;Wu Hongming

机构地区：[1]武汉理工大学物流工程学院,武汉430063

出　　处：《起重运输机械》2021年第21期50-55,共6页Hoisting and Conveying Machinery

摘　　要：拣选作业是物流中心最耗费人力和时间的过程,采用高效的拣选方式对缩短物流时间、降低物流成本有着重要的影响。文中针对目前发展最为迅速的货到人拣选方式,提出了人机协作拣选工作台方法,使用拣选机械臂和人类共同在工作台进行协作拣选。人机协作拣选方法与人工拣选方法相比降低了人类劳动强度,提高工作效率;与全自动拣选方法相比降低了拣选系统的复杂度,对不同场景的适用性更好。为了实现拣选工作台中的人机协作,使用Kinect传感器提取人体动作特征,并利用支持向量机方法对特征进行分类,实现拣选时的人体动作识别,使拣选机械臂理解人类行为,从而更加顺畅地与人类合作。最终,在测试集上的识别准确率达到了94.3%,为后续实现拣选工作台中的人机协作奠定了基础。Since picking is the most labor-intensive and time-consuming work in logistics centers,it is important to adopt efficient picking methods to shorten logistics time and reduce logistics costs.For the“goods-to-person”picking method,a method developing rapidly at present,a man-machine cooperative picking method is proposed,in which the picking robot arm and human cooperate in picking on the workbench at the same time.Compared with manual sorting method,man-machine collaborative sorting reduces human labor intensity and improves work efficiency.Compared with the automatic picking method,it reduces the complexity of the picking system and has better applicability to different scenes.In order to realize human-machine cooperation,firstly,A Kinect sensor is used to extract human motion features,and support vector machine is used to classify the features and identify the human motion during picking,so that the picking robot arm can understand human behavior and cooperate with humans more smoothly.Finally,the recognition accuracy on the test set reaches 94.3%,which laid a foundation for the future realization of human-computer cooperation.

关 键 词：货到人拣选 人机协作 拣选工作台 动作识别 支持向量机 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置]