ERI: Adaptive and Resilient Communication-Aware Multi-Robot Coordination

这项工程研究启动（ERI）拨款将支持基础研究，旨在提高在现实和可能的对抗环境中运行的自主移动的机器人团队的可靠协调。能够协作的机器人在从搜索和救援任务到精准农业的许多应用中显示出巨大的潜力。由于有限的感知、通信和处理能力，协作组中的自主机器人通常需要通过邻近的通信来做出集体决策。例如，通过范围受限的无线通信与相邻机器人交换传感数据或控制命令。无线通信的不可或缺的作用要求机器人除了完成原有任务外，还能对不断变化的环境进行协调和反应，以实现可靠的信息共享。然而，现有的研究范式往往假设通信环境是很好的建模，可以预先编程到机器人系统的有效协调，而不会中断。这极大地限制了机器人团队的能力，并使它们在通常不可预测的现实环境中变得脆弱。为了应对这些挑战，该项目旨在创建一套方法和工具，通过使机器人能够在飞行中学习环境，使其运动适应环境变化，并在发生意外机器人故障时保持通信，来增强许多多机器人协调任务。该项目还将支持教育和外展活动，如课程开发，扩大代表性不足的群体的学生的参与，以及当地社区参与夏洛特大都市区。该项目的目标是创建新的方法和算法，使学习，通信和运动控制的移动的机器人团队的共同设计。这可以允许机器人的鲁棒操作，其具有对通信能力的可证明保证和适应于不确定环境的多机器人网络弹性，积极支持主要任务执行。为实现这一目标，该项目将作出两项主要贡献：（i）开发数据驱动的方法，使机器人在线协作学习空间变化的现实通信性能，以及（ii）开发新方法，指定现实通信约束和网络弹性对任务相关机器人运动的影响，以实现联合优化，从而在保证性能的情况下实现更有效的多机器人协调。这项工作将在物理机器人平台上的模拟和实验中进行评估。该项目由跨部门的机器人基础研究项目支持，该项目由工程部（ENG）和计算机与信息科学与工程部（CISE）共同管理和资助。该奖项反映了NSF的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

This Engineering Research Initiation (ERI) grant will support fundamental research that aims to improve reliable coordination for a team of autonomous mobile robots operating in realistic and possibly adversarial environments. Robots that can collaborate have shown great potential in many applications from search and rescue missions to precision agriculture. Due to limited sensing, communication, and processing capabilities, autonomous robots in a collaborating group often need to make collective decisions through communications in proximity. For example, exchanging sensing data or control commands with neighboring robots through range-constrained wireless communications. The integral role of wireless communications requires robots to coordinate and react to the changing environment for reliable information sharing in addition to their original tasks. However, existing research paradigm often assumes that the communication environment is well modeled and can be pre-programmed into robotic systems for effective coordination without disruptions. This significantly limits the capabilities of robot teams and makes them vulnerable in real-world environments that can often be unpredictable. To address the challenges, this project seeks to create a set of methods and tools that may augment many multi-robot coordination tasks, by enabling robots to learn the environment on the fly, adapt their motion to environmental changes, and maintain communications when unexpected robot failures happen. The project will also support education and outreach activities such as curriculum development, broadening participation of students from underrepresented groups, and local community engagement in the Charlotte metropolitan area.The objective of this project is to create novel methods and algorithms that enable the co-design of learning, communication, and motion control for mobile robot teams. This may allow for robust operation of robots with provable assurances on communication capabilities and multi-robot network resilience adaptive to uncertain environments, positively supporting the primary task execution. In pursuit of this goal, the project will make two main contributions: (i) Developing data-driven methods for robots to collaboratively learn the spatially varying realistic communication performance online, and (ii) Developing new approaches that specify the impact of realistic communication constraints and network resilience on the task-related robots’ motion, to enable joint optimization for more efficient multi-robot coordination with performance guarantees. The work will be evaluated in simulations and experiments on physical robotic platforms. This project is supported by the cross-directorate Foundational Research in Robotics program, jointly managed and funded by the Directorates for Engineering (ENG) and Computer and Information Science and Engineering (CISE).This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.