FRR: Collaborative Research: Collaborative Learning for Multi-robot Systems with Model-enabled Privacy Protection and Safety Supervision

该项目为多机器人系统提供了一种协作强化学习方法，以确保安全和隐私保护。该方法使机器人能够在系统约束下不断学习和适应动态情况。此外，这种新方法确保了参与机器人的身份和位置等隐私信息在涉及多个参与机器人的协作任务中得到保护。虽然多机器人学习的方法是通用的，但它可以应用于由自主和半自动汽车和卡车驱动的道路上的智能交通系统。这种方法的演示将在一个真实的城市环境的全面测试环境中进行。该项目将基于模型的安全性与无模型强化学习相结合，使强化学习适用于安全关键型协作多机器人系统。它将首先使用基于深度koopman的一般非线性机器人系统安全调节来解决单机器人强化学习问题，以保证安全同时保持学习效率。然后将结果扩展到多机器人集体强化学习，其中机器人部署在共享，竞争或资源受限的环境中。通过利用协作学习的内在动态，该项目还将为学习过程中收集和共享的数据提供基于动态的隐私保护。与传统的隐私机制（要么以准确性换取隐私，要么导致沉重的计算/通信开销）不同，启用动态的隐私方法可以在产生很少的计算/通信开销的同时保持学习的最优性。算法和框架将通过数值模拟和真实连接车辆在真实轨道上的实验进行评估。该项目的成果将用于丰富研究生和本科课程。pi还将利用现有的各种正在进行的外展机会，激发K-12学生和社区大学技术人员对STEM的兴趣。该项目由跨部门机器人基础研究项目支持，由工程（ENG）和计算机与信息科学与工程（CISE）联合管理和资助。该奖项反映了美国国家科学基金会的法定使命，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。

This project provides a collaborative reinforcement learning approach for multi-robot systems that ensures safety and is privacy-preserving. The approach enables robots to continuously learn and adapt to dynamic situations within the systems constraints. Moreover, this new approach ensures that the private information of the participating robots, such as identity and position, can be protected in collaborative tasks involving multiple participating robots. While the approach for multi-robot learning is general, it has application to intelligent transportation systems on roadways driven by autonomous and semi-autonomous cars and trucks. A demonstration of this approach is to be conducted in a full-scale test environment of a realistic urban setting.This project combines model-based safety with model-free reinforcement learning to enable reinforcement learning's applicability to safety-critical collaborative multi-robot systems. It will first address single-robot reinforcement learning using a deep Koopman-based safety regulation for general nonlinear robotic systems to guarantee safety while retaining learning efficiency. The result will then be extended to multi-robot collective reinforcement learning where robots are deployed in shared, contested, or resource-constrained environments. By exploiting the inherent dynamics of collaborative learning, the project will also enable dynamics-based privacy protection for collected and shared data during learning. Different from conventional privacy mechanisms that either trade accuracy for privacy or incur heavy computation/communication overhead, the dynamics-enabled privacy approach can maintain learning optimality while incurring little computation/communication overhead. The algorithms and frameworks will be evaluated using both numerical simulations and experiments with real connected vehicles on real tracks. Results of the project will be used to enrich both graduate and undergraduate courses. The PIs will also use existing various on-going outreach opportunities to energize interests in STEM in K-12 students and community college technicians.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.