Research on Multibody Dynamics and Control for Collaborative Elastic Object Transportation by a Heterogeneous Swarm with Aerial and Land-Based Mobile Robots

该研究项目研究了一群不同种类的机器人协同运输软物体的问题，这些机器人既包括地面机器人，也包括空中机器人。今天，移动机器人已经被用来完成重要的任务。然而，不是使用单个机器人，而是协作的一群机器人可以在性能、对单个机器人故障的稳健性和灵活性方面具有固有的优势。异质群体带来了更多可能的优势，例如，陆基机器人可能能够非常稳健和高效地在环境中运输物体，但空中机器人可能需要克服否则无法克服的障碍，让位于合作的运输方案。在其新颖性方面，本项目提出了各种具有挑战性的研究机会，包括作为一个整体的群体协调、单个机器人的稳健推进控制以及能够准确和合作地操纵软对象的机器人机械手的控制。这些严峻的挑战需要新的方法，因此，对先进控制理论概念的研究和仿真测试，包括分布式和信息论模型预测控制。从此以后，该项目的目标是促进一套关于控制机器人群的扩展方法。为此，该项目包括为所考虑的不同种类的机器人群建立数学模型和开发一个包罗万象的控制和仿真框架。这使得协同运输任务的仿真研究成为可能，此外，还包括运输对象从陆基到空中机器人的移交，反之亦然。同时，为了验证所研究的方法在仿真中被认为是有前途的，该项目的目标是开发和进行与所开发的整体仿真和控制环境紧密一致的硬件实验。这涉及到设计合适的硬件机器人和装配机械手，以及研究合适的传感、定位、导航和通信方法。相关研究小组之间的密切合作，以及他们在研究真实世界和定制的机器人硬件以及机械系统的控制和仿真方面的专业知识，使弥合控制理论、模拟和真实世界实验之间的差距成为可能，而这往往是群体机器人研究中的一个问题。从数学模型、控制理论和仿真一直到彻底的硬件实验，所进行的研究可以假设一个前所未有的范围和深度。

This research project investigates the collaborative transportation of soft objects by a heterogeneous swarm of robots involving land-based as well as aerial robots. Already today, mobile robots are used to accomplish important tasks. However, not employing a single robot, but rather a cooperating swarm of robots can have inherent advantages with respect to performance, robustness to individual robot failures, and flexibility. A heterogeneous swarm brings about even more possible advantages, e.g., land-based robots may be able to transport an object very robustly and efficiently through an environment, but an aerial robot might be necessary to overcome otherwise insurmountable obstacles, giving way to cooperative transportation schemes. In its novelty, the present project poses various challenging research opportunities, including the coordination of the swarm as a whole, the robust propulsion control of the individual robots as well as the control of the robots' manipulators for being able to accurately and cooperatively manipulate soft objects. Those demanding challenges necessitate novel approaches and, therefore, the research and simulative testing of advanced control theoretic concepts, including distributed and information-theoretic model predictive control. Henceforth, the project aims to contribute to an extended set of methodology with respect to the control of robot swarms. To this end, the project includes the mathematical modelling and development of an encompassing control and simulation framework for the considered heterogeneous robot swarm. This enables the simulative research of cooperative transportation tasks, including additionally the hand-over of the transported object from land-based to aerial robots and vice versa.In parallel, to validate the researched approaches deemed promising in simulations, the project aims to develop and conduct hardware experiments in close concordance with the developed integral simulation and control environment. This involves the design of appropriate hardware robots and fitting manipulators as well as the research on suitable methods for sensing, localization, navigation and communication. The close collaboration between the involved research groups, with their expertise in research with real-world and custom-made robotic hardware as well as the control and simulation of mechanical systems, makes it possible to bridge the gap between control theory, simulation and real-world experiments that is often an issue in swarm robotics research. The undertaken research can assume an otherwise unprecedented scope and depth, leading from the mathematical model, control theory and simulation all the way to thorough hardware experiments.