面向公共卫生事件应急处置的防疫机器人环境深度感知方法研究

防疫机器人辅助医护人员应对公共卫生应急事件（如新冠病毒）处置，进行消毒、转运和协作“抗疫”具有不可替作用，但其实际应用成效不高。本项目针对其应用背景，分析影响其处置规划和作业控制成因，是由于疫情环境高动态性和不确定性与机器人环境感知能力不足的科学问题。为此本项目提出针对疫情环境建立的“目标-患者-度量、热场地图”深层系统感知的科学思想，建立三通道防疫目标学习模型，提出跨模态患者检测与定位方法，并设计环境度量、热场模型，多个层面提升防疫机器人环境感知技术的稳定性与适应能力。本项目从原理和方法上探索机器人对防疫未知环境感知多元信息建模与感知方法，突破疫情环境动态、不确性和复杂等条件制约，为支撑防疫机器人全局规划与移动作业提供基础理论与系统方案。本项目从公共卫生应急环境实际需求凝练关键科学问题，促进基础成果走向应用，对推进防疫机器人快速部署，助力阻断疫情发展具有重要实际意义和科学价值。

For the treatment of public health emergency events (such as COVID-19), anti-epidemic robots can have an irreplaceable role of helping medical staff in disinfection, transportation and collaboration, but its actual application is not effective. Based on such application background, this project analyzes the causes that affect its disposal planning and operation control. It is a scientific problem due to the high dynamics and uncertainty of the epidemic environment and the weakness of robot environmental perception. To this end, the project proposes a deep perception scientific idea of "target-patient-metric, thermal field map" established for the epidemic environment. In the project, a three-channel anti-epidemic target learning model is establish and a method of cross-modal patient detection and localization is proposed. Those contributions will improve stability and adaptability of the robot perception technology in anti-epidemic environment at multiple levels. The project will explore the method of robots' multi-information modeling and perception for epidemic-preventing unknown environment from the principles and methods, break through the bottleneck of robotic perception technology under the conditions of epidemic environment dynamics, uncertainty and complexity, so it will provide solid foundation theory and platform technical support for the global planning and mobile operations of epidemic prevention robots. Finally, results of such research can promote basic achievements to application. It has important practical significance and scientific value for promoting rapid deployment of epidemic prevention robots and helping to stop the development of epidemic outbreaks.