Large-scale Multi-robot System

大型多机器人系统

• 批准号: RTI-2021-00527
• 负责人: Saeedi, Sajad
• 金额: $ 2.85万
• 依托单位: Ryerson University
• 依托单位国家: 加拿大
• 项目类别: Research Tools and Instruments
• 财政年份: 2020
• 资助国家: 加拿大
• 起止时间: 2020-01-01 至 2021-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=717867
• 关键词: Large; scale; Multi; robot; System

Multi-robot systems have benefits such as being more robust to failures and being faster in accomplishing the tasks. In a distributed and multi-agent system, the whole team is more robust since the failure of one of the robots does not halt the entire mission. In applications such as surveying large areas or deploying first responders in natural disasters, multi-robots systems are able to achieve the objectives faster and more efficiently. Topics such as task allocation, formation control, and coordination in multi-robot systems are being actively investigated by researchers. With the advent of the recent machine learning and deep learning methods, promising results are being presented to address the existing problems such as time and space complexity in multi-agent systems. One of the key areas in multi-agent systems is multiple-robot simultaneous localization and mapping (SLAM), motivated by the fact that exploration and mapping tasks can be done faster and more accurately by multiple robots than by a single robot. Many collaboration-based operations need to be completed fast and autonomously and require localization and mapping. Some of these applications include fire fighting, cleaning operations (like removing marine oil spills), underwater exploration, rescue operations, and maintenance investigations. An important advantage of team-based SLAM is that different types of robots, equipped with different sensors, can provide a better understanding of the environment. The proposed equipment, a large-scale heterogeneous team of robots, is essential for research in multiple-robot SLAM and also for research in other programs such as formation control, task allocation, etc. The proposed equipment is urgently needed to be used in novel research programs such as using haptics and manipulation-based SLAM in multiple-robot SLAM, multi-robot active neural SLAM, developing multi-robot disinfection systems to prevent the spread of COVID-19, and improving the computational complexity of multi-robot systems using machine learning methods. The proposed equipment will be used to verify the algorithmic designs in real-world scenarios with dynamic and moving objects.

多机器人系统具有诸如对故障更加鲁棒并且能够更快地完成任务等优点。在分布式多代理系统中，整个团队更加稳健，因为其中一个机器人的故障不会停止整个任务。在勘察大面积区域或在自然灾害中部署急救人员等应用中，多机器人系统能够更快、更高效地实现目标。研究人员正在积极研究多机器人系统中的任务分配、编队控制和协调等主题。随着最近机器学习和深度学习方法的出现，人们正在提出有希望的结果来解决多智能体系统中存在的时间和空间复杂性等问题。 多智能体系统的关键领域之一是多机器人同时定位和建图（SLAM），其动机是多个机器人比单个机器人可以更快、更准确地完成探索和建图任务。许多基于协作的操作需要快速、自主地完成，并且需要本地化和绘图。其中一些应用包括消防、清洁作业（如清除海洋石油泄漏）、水下勘探、救援作业和维护调查。基于团队的 SLAM 的一个重要优势是，不同类型的机器人，配备不同的传感器，可以更好地了解环境。 所提出的设备是一个大规模异构机器人团队，对于多机器人 SLAM 的研究以及编队控制、任务分配等其他项目的研究至关重要。所提出的设备迫切需要用于新颖的研究项目，例如在多机器人 SLAM 中使用基于触觉和操作的 SLAM、多机器人主动神经 SLAM、开发多机器人消毒系统以防止病毒传播。 COVID-19，并使用机器学习方法提高多机器人系统的计算复杂性。所提出的设备将用于验证具有动态和移动物体的现实场景中的算法设计。