Human-Drones Teaming for Autonomous Rescue with Fast Reinforcement Learning

人机协作通过快速强化学习进行自主救援

• 批准号: 2788094
• 金额: --
• 依托单位: CRANFIELD UNIVERSITY
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2023
• 资助国家: 英国
• 起止时间: 2023 至 无数据
• 项目状态: 未结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-2788094
• 关键词: Human; Drones; Teaming; Autonomous; Rescue

In many emerging crisis areas, the environment has changed compared to prior knowledge. This makes autonomous tasks such as search and rescue very challenging, as it cannot exploit prior maps and contextual knowledge. Human-drone teaming (HDT) can overcome some of these challenges and requires a mutual understanding between the human ground stakeholders and drones. It is paramount to study how autonomous drones can reason the specific situation efficiently and improve the situation awareness based on the prior knowledge from the human experts to accomplish the rescue task successfully.

在许多新出现的危机地区，与先前的知识相比，环境发生了变化。这使得搜索和救援等自主任务非常具有挑战性，因为它无法利用先前的地图和背景知识。人-无人机合作(HDT)可以克服其中一些挑战，需要人类地面利益攸关方和无人机之间的相互理解。研究自主无人机如何基于人类专家的先验知识，有效地进行特定情况的推理，提高态势意识，对于顺利完成救援任务具有重要意义。