NRI: Small: Multirobot-Human Coordination for Visual Scene Understanding

这项研究的目标是开发配备视觉和其他传感器的机器人团队，能够在关键任务中与人类一起工作，如搜索和救援。关键要求是了解情况和协调行动。方法是开发数学框架和算法，使这样的机器人团队能够协调它们的路径，共享和分析它们的传感器数据，维护通信，并有效和安全地与人类互动。该项目汇集了计算机视觉、机器人、评估理论和控制方面的专家。实现上述目标需要在几个相互关联的领域取得进展。具体地说，传感、估计和轨迹控制任务必须无缝地将视觉分析与导航和控制策略以及来自人类的输入相结合。必须开发新的分布式评估策略，以适应困难和动态的环境。高效的人-机器人协调需要联合探索和测绘的方法，识别重要的视觉信息，以及机器人？在不同的自主水平下运行。更大的影响。该项目的成功将是向部署机器人团队迈出的重要一步，以帮助人类执行危险和复杂的任务，如灾难应对。搜救专家将建议团队开发一个原型系统，并在模拟操作条件的情况下对其进行评估。开发的软件工具将传播给其他研究人员，这样他们就可以在成果的基础上再接再厉。来自UCR高度多样化的学生群体的本科生将获得与研究生研究人员一起工作的宝贵经验。

The objective of this research is to enable the development of teams of robots, equipped with vision and other sensors, capable of working alongside humans in critical missions, such as search and rescue. Key requirements are situational awareness and coordinated action. The approach is to develop mathematical frameworks and algorithms to enable such a team of robots to coordinate their paths, share and analyze their sensor data, maintain communications, and interact effectively and safely with humans. The project brings together experts in computer vision, robotics, estimation theory and controls.Intellectual Merit. Realizing the above goals will require advances in several inter-related domains. Specifically, the sensing, estimation, and trajectory control tasks must seamlessly integrate visual analysis with navigation and control strategies, as well as inputs from humans. Novel distributed estimation strategies must be developed to accommodate difficult and dynamic environments. Efficient human-robot coordination necessitates methodologies for joint exploration and mapping, identifying important visual information, and robots? operation at different levels of autonomy.Broader Impact. The success of this project will be a major step towards the deployment of teams of robots to assist humans in dangerous and complex tasks like disaster response. Search-and-rescue experts will advise the team in developing a prototype system, and evaluating it in situations that mimic operational conditions. The developed software tools will be disseminated to other researchers so they can build on the results. Undergraduates from UCR's highly diverse student population will gain valuable experience working alongside graduate student researchers.