首页|Studies from Shanghai Jiao Tong University Provide New Data on Robotics (A Robot ic Manipulation Framework for Industrial Human-robot Collaboration Based On Cont inual Knowledge Graph Embedding)
Studies from Shanghai Jiao Tong University Provide New Data on Robotics (A Robot ic Manipulation Framework for Industrial Human-robot Collaboration Based On Cont inual Knowledge Graph Embedding)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News – Current study results on Robotics have been published. According to news reporting out of Shanghai, People’s Republic of China, by NewsRx editors, research stated, “Hybrid robots can assist human wo rkers in various tasks due to their integration of mobility and manipulability. The rapid diffusion of these robots in factories has significantly elevated the automation and intelligence level of manufacturing, while also brings challenges to human-robot collaboration.” Financial support for this research came from National Key R&D Prog ram of China. Our news journalists obtained a quote from the research from Shanghai Jiao Tong University, “Traditionally, human workers need to instruct robots to perform a r ange of tasks by explicitly demonstrating these operations. However, this proces s imposes excessive burdens on workers as the tasks and environment for robots b ecome more and more diversified and complex. To alleviate this issue, we propose an innovative robotic manipulation framework based on continual knowledge graph embedding. This framework enables hybrid robots to break free from the constrai nts of fixed rules set by human demonstrations, instead endowing them with infer ring capability. The core idea is to utilize semantic information related to obj ects (such as category, material, and components) and tasks assigned to infer ap propriate operational parameters for robots via a knowledge graph. These operati onal parameters include the suitable type of gripper, the proper area for object manipulation, and the reasonable force range for effective grasping. We conduct an experimental analysis of the proposed framework with a real-world hybrid rob ot, which performed 158 different tasks involving 46 objects commonly seen in in dustry, achieving a success rate of up to 96.8%. Furthermore, our f ramework can continuously enhance the adaptability of robotic operations and eff ectively balance the learning of new and old knowledge.”
ShanghaiPeople’s Republic of ChinaAs iaEmerging TechnologiesMachine LearningNano-robotRobotRoboticsRobotsShanghai Jiao Tong University
NETL
NSTL
