CAREER: Controlling Ecologically Destructive Processes with a Network of Intelligent Robotic Agents

该项目旨在通过一组机器人的智能、协调干预来控制破坏性的环境过程，如森林火灾、石油泄漏和农业病虫害。这需要开发全新的控制理论和算法工具，以驱动机器人采取控制行动来调节环境过程。机器人的控制行为围绕着机器人和环境形成了一个大规模的反馈环，导致了复杂的动力学现象。提出了机器人-环境耦合系统在三种不同时间尺度下的控制策略：(I)与机器人动力学相比，环境变化较慢；(Ii)环境与机器人动力学在同一时间尺度上，寻求即时控制效果；(Iii)环境与机器人动力学在同一时间尺度上，寻求长期控制效果。提出了三种不同的基于优化的技术来生成每个区域的分散控制策略。研究了在这些策略下机器人-环境系统的稳定性、收敛和最优性。此外，在实验室和室外用四旋翼空中机器人网络进行的实验证明了控制策略的实用性。该项目还纳入了一个全面的教育和推广计划，使用四旋翼机器人作为教学工具，接触到所有年级不同背景的学生。最终，该项目寻求通过为新的机器人技术奠定基础，减轻石油和其他化学品泄漏、森林火灾、虫害和其他生态破坏现象造成的经济、社会和生态破坏。

This project aims to control destructive environmental processes such as forest fires, oil spills, and agricultural pest infestations through the intelligent, coordinated intervention of a group of robots. This requires the development of fundamentally new control theoretic and algorithmic tools to drive the robots to take control actions to regulate the environmental process. The robots' control actions close a large-scale feedback loop around the robots and the environment, giving rise to complex dynamical phenomena. The project proposes control strategies for this coupled robot-environment system in three different timescale regimes: (i) the environment changes slowly compared to the robots' dynamics, (ii) the environment and robot dynamics are on the same timescale and immediate control effect is sought, and (iii) the environment and robot dynamics are on the same timescale and long-term control effect is sought. Three different optimization based techniques are proposed to generate decentralized control strategies for each regime. Stability, convergence, and optimality properties of the robot-environment system under these strategies are studied. Furthermore, experiments with a network of quadrotor aerial robots, both in the lab and outdoors, demonstrate the practicality of the control strategies. The project also incorporates a comprehensive education and outreach program using quadrotor robots as teaching tools to reach students from diverse backgrounds at all grade levels. Ultimately, the project seeks to alleviate the economic, social, and ecological damage caused by oil and other chemical spills, forest fires, pest infestations, and other ecologically destructive phenomena by laying the foundations of a new robotic technology.