CAREER: Robots that Plan Interactions, Come and Go, and Build Trust

经过几十年的根本性进步，我们开始看到自主机器人对关键的社会和经济问题产生有意义的影响。 虽然非常令人兴奋，但这些进步主要出现在单机器人系统中，鉴于多机器人系统的高度复杂性，这是一个自然的进展。为了扩展到现实世界的问题，这个教师早期职业发展（CAREER）项目推进了多机器人系统理论，允许机器人智能地规划他们的交互，优雅地进入和退出系统，并与其他机器人和人类队友一起参与可信的决策过程。 此外，本项目的理论工作也适用于多机器人多人搜救。 事实上，当机器人和人类在大荒野中寻找迷路的人时，上面列出的理论进步将被证明是无价的。 最后，该项目包括一个全面的教育和推广计划，包括：跨部门自主性课程;以残疾人为重点的教学计划;为工程专业代表性不足的学生提供K-12学术经验;以及国家和国际教育和推广渠道。一个基于独立系统的框架，用于随着时间的推移用采样和基于梯度的方法来建模机器人交互结构，以有效地计算交互计划;一个组合优化框架，用于建模最优开放的多机器人系统，产生机器人进入和退出系统的计划，同时保证底层协作目标的正确性;一个框架的信任建立在协作多机器人多人决策的基础上多武装土匪，我们称之为可信的多武装土匪的概念;一组搜索和救援案例研究，用于评估我们的研究推力;和一个便携式，室内/室外，多尺度的实验平台，用于实验验证异构多机器人团队。 该项目解决了对自主协调的灵活性、可扩展性和弹性至关重要的四个领域的基本挑战：（1）以适应高级别使命目标和部署环境的方式规划其交互的系统，同时尊重低级别协作要求;（2）组成随时间变化的系统，同时保持对此类变化的弹性;（3）随着时间的推移，从其他系统中选择积极建立信任的动作的系统;以及（4）在不同部署规模的现实条件下原型化和测试的系统。该项目由跨部门机器人基础研究计划支持，由工程局（ENG）和计算机与信息科学与工程局（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知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

After decades of fundamental advances, we are beginning to see meaningful influence of autonomous robots on crucial social and economic problems. While tremendously exciting, these advances are primarily seen in single robot systems, a natural progression given the highly complex nature of multi-robot systems. In order to scale to real-world problems, this Faculty Early Career Development (CAREER) project advances multi-robot system theory to allow robots to plan their interactions intelligently, gracefully enter and exit systems, and participate in trustful decision-making processes with other robots and human teammates. In addition, the theoretical work in this project applies to multi-robot multi-human search and rescue. Indeed, when robots and humans search for a lost person in large wilderness, the theoretical advancements listed above will prove invaluable. Finally, this project includes a comprehensive education and outreach plan consisting of: curriculum focused on autonomy that crosses departmental boundaries; pedagogical programs with an emphasis on persons with a disability; K-12 academic experiences for underrepresented students in engineering; and channels for national and international education and outreach.This project focuses on developing new theory and technologies, including: a framework based on independence systems for modeling robot interaction structures over time with sampling and gradient-based methods for efficiently computing interaction plans; a combinatorial optimization framework for modeling optimally open multi-robot systems yielding plans for robots entering and exiting systems while guaranteeing the correctness of underlying collaborative objectives; a framework for trust-building in collaborative multi-robot multi-human decision-making based on multi-armed bandits, a concept we call the trustful multi-armed bandit; a set of search and rescue case studies for evaluating our research thrusts; and a portable, indoor/outdoor, multi-scale testbed for experimental validation of heterogeneous multi-robot teams. This project addresses fundamental challenges in four areas critical to the flexibility, scalability, and resilience of autonomous coordination: (1) systems that plan their interactions in a manner that adapts to high-level mission objectives and the deployment environment, while respecting low-level collaboration requirements; (2) systems whose composition changes over time while remaining resilient to such changes; (3) systems that select actions that actively build trust from other systems over time; and (4) systems that are prototyped and tested under realistic conditions across varying scales of deployment.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.