AerialHitches: Forming and Controlling Hitches for Fully Autonomous Transportation Using Aerial Robots with Cables

该项目将改进空中机器人，使它们能够在没有人为干预的情况下更好地抓取、操纵和运输物体。传统的空中机器人依靠抓手（即末端执行器）和机械臂来操纵和运输物体。这些部件不仅大大增加了重量，而且减少了飞行时间，使机器人的控制复杂化。该项目引入了一种创新的替代方案，其中机器人利用轻质电缆形成“连接线”来执行操作任务，绕过了对沉重的外部组件的需求。电缆的优点是重量轻且灵活，可以实现传统刚性夹具无法提供的多功能操作技术。电缆还可以适应各种形状和大小的物体，增加了空中机器人可以执行的任务范围。此外，电缆的使用大大降低了系统的整体重量，从而增加了飞行时间。这一突破性的概念有可能极大地提高建筑、物流和灾害管理等领域的生产率，这些领域的高频、重复运输任务很常见。在这些领域，使用基于电缆的空中机器人系统可以降低人工运输和在高架或危险区域工作的安全风险。在技术方面，该项目需要将常见的电缆配置抽象到有限维空间中，以管理电缆的物理动态。强化学习算法将在此抽象的基础上制定，利用理想动力学模型加速收敛到最优行为。将开发基于拓扑的规划算法来形成所需的特定运动。代替从零开始制定计划，将建立一个故障库来简化运输过程，可以根据运输需求进行调整。通过这个项目，潜在的贡献包括在控制交错电缆的强化学习方面的进步，在空中机器人中使用拓扑框架对电缆建模，以及推进我们对控制空中机器人行为和操作原理的理解。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 will improve aerial robots so they can grasp, manipulate and transport objects better and without human intervention. Conventional aerial robots rely on grippers (i.e., end effectors) and robotic arms for manipulating and transporting objects. These parts not only add significant weight, but also reduce flight duration and complicate the control of the robots. This project introduces an innovative alternative, wherein the robots utilize lightweight cables to form 'hitches' for performing manipulation tasks, bypassing the need for weighty external components. The benefit of cables is that they are lightweight and flexible, allowing for versatile manipulation techniques that traditional rigid grippers cannot provide. Cables can also conform to objects of various shapes and sizes, increasing the range of tasks an aerial robot can perform. Additionally, using cables dramatically reduces the overall weight of the system, resulting in increased flight duration. This groundbreaking concept has the potential to greatly improve productivity in fields such as construction, logistics, and disaster management, where high-frequency, repetitive transportation tasks are common. In these sectors, using a cable-based aerial robotic system could reduce safety risks associated with manual transportation and work in elevated or hazardous areas.On a technical front, the project entails abstracting common cable configurations into a finite-dimensional space to manage the physical dynamics of the cables. A reinforcement learning algorithm will be formulated based on this abstraction, expediting the convergence to optimal actions by utilizing an ideal dynamics model. Topology-based planning algorithms will be developed to form the hitches, accounting for the specific movements required. In lieu of creating plans from scratch, a library of hitches will be established to streamline the transportation process, which can be adjusted according to transportation requirements. Through this project, the potential contributions include advancements in reinforcement learning for controlling interlaced cables, modeling cables using topological frameworks in aerial robotics, and progressing our understanding of the principles governing aerial robot behavior and manipulation.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.