首页|Studies from Shandong University Have Provided New Information about Robotics (H ogn-tvgn: Human-inspired Embodied Object Goal Navigation Based On Time-varying K nowledge Graph Inference Networks for Robots)
Studies from Shandong University Have Provided New Information about Robotics (H ogn-tvgn: Human-inspired Embodied Object Goal Navigation Based On Time-varying K nowledge Graph Inference Networks for Robots)
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
Investigators discuss new findings in Robotics. According to news reporting originating in Weihai, People's Republic o f China, by NewsRx journalists, research stated, "Object goal navigation tasks a re critical for robots operating in unfamiliar environments, where they must loc ate specific objects using visual cues. The ability to leverage prior knowledge significantly enhances a robot's associative capabilities, leading to improved n avigation performance." Financial supporters for this research include National Natural Science Foundati on of China (NSFC), Natural Science Foundation of Shandong Province, China Postd octoral Science Foundation, Young Scholars Program of Shandong University, Weiha i. The news reporters obtained a quote from the research from Shandong University, "However, existing methods struggle with the generalization challenge when trans ferring navigation models to new environments, a key issue addressed in this pap er. To overcome this challenge, on the one hand, a time-varying knowledge graph is proposed to update the prior knowledge graph with context vectors derived fro m co-occurrence objects in the current observation. This approach prioritizes lo cal graphs centered around the target and co-occurring objects, allowing for eff icient and accurate target localization. Furthermore, the dynamic updating mecha nism facilitates efficient exploration in new scenarios. On the other hand, to e mbed prior knowledge more rationally in the reinforcement learning-based navigat ion strategy, a timevarying knowledge graph inference network (TVGN) is present ed. The TVGN utilizes context vectors and global spatial semantic information to perceive and understand the environment in real-time. It formulates navigation strategies based on the precise goal information encoded within the graph, there by enhancing the robot's efficiency in reaching the target. Based on the widely applied dataset AI2-THOR, extensive comparative experiments are conducted to ill ustrate the effectiveness of the proposed method."
WeihaiPeople's Republic of ChinaAsiaEmerging TechnologiesMachine LearningNano-robotRobotRoboticsShandong University
NETL
NSTL
