A technique that allows a formation-enforcing control (FEC) derived from graph rigidity theory to interface with a realistic relative localization system is proposed in this paper. Recent research in sensor-based multi-robot control has given rise to multiple modalities of mutual relative localization systems. In particular, vision-based relative localization has reached the stage where it can be carried onboard lightweight UAVs in order to retrieve the relative positions and relative orientations of cooperating units. A separate stream of development spawned distributed formation-enforcing control which can lead individual robots into a desired formation using relative localization of their neighbors. These two fields naturally complement each other by achieving real-world flights of UAVs in formation without the need for absolute localization in the world. However, real relative localization systems are, without exception, burdened by non-negligible sensory noise, which is typically not fully taken into account in formation-enforcing control algorithms. Such noise can lead to rapid changes in velocity, which further interferes with visual localization. Our approach provides a solution to these challenges, enabling practical deployment of FEC under realistic conditions, as we demonstrated in real-world experiments.
翻译:本文提出了一种技术,使得从图刚性理论导出的编队控制(FEC)能够与现实中的相对定位系统相接口。最近传感器为基础的多机器人控制的研究出现了多种互相相对的定位系统模态。特别是,基于视觉的相对定位已经发展到可以在轻型无人机上执行,以检索合作单元的相对位置和相对方向。分布式编队控制产生了一个分开的发展流,在使用邻居的相对定位将单个机器人引导到所需编队方面取得了进展。这两个领域自然地相互补充,通过在世界中不需要绝对定位实现无人机集体飞行。然而,真实的相对定位系统无一例外地受到不可忽略的传感器噪声的影响,通常在编队控制算法中并没有完全考虑。这些噪声可能导致速度的快速变化,进一步干扰视觉定位。我们的方法提供了解决这些挑战的解决方案,在真实条件下实现了FEC的实用部署,正如我们在实际操作中所证明的那样。