Optimization of Motion Planning and Navigation for Aerial Vehicles with Differential Constraints through Analysis and Control of its Observability

在未知环境中运行的基于图像的单摄像头导航系统必须解决基本的缩放问题。这意味着，不可能区分远距离和快速移动的物体以及移动缓慢的近距离物体。在某些情况下，附加的惯性感应可以解决这种模糊性。确定解决这种模糊性的相关条件是第一个项目阶段的一个主要目标。在运动规划中考虑了这些条件，解决了定标问题，从而系统地提高了导航性能。为了确定位置，除了伸缩问题外，还需要飞行器的再访问能力。允许重复访问的方法在文献中是可用的，然而，这些方法主要是基于启发式的。缺少一种理论上的处理方法，该方法为重新访问循环提供了系统的程序(即，何时需要它们以及需要它们的程度)。这项提议旨在缩小这一差距。为此，应推导出可实现的导航估计性能的通用度量。在此基础上，发展了一种考虑固定翼飞行器差动约束的包括位置在内的综合运动规划和导航的理论方法。

Image-based navigation systems with mono-cameras that are operated in unknown environments have to cope with a fundamental scaling problem. This means, it is not possible to distinguish between far distant and fast moving objects and close objects that are moving slowly. In some situations, additional inertial sensing allows to solve this ambiguity. The determination of the associated conditions for such an ambiguity resolution was one major objective of the first project phase. These conditions were considered in the motion planning, which allowed a solution of the scaling problem and thus, a systematic improvement of the navigation performance. In order to determine the position in addition to the scaling problem, a re-visiting capability of the aerial vehicle is required. Methods that allow such re-visits are available in literature, however, these are mainly based on heuristics. A theoretical treatment that provides a systematic procedure for re-visiting loops (i.e. when and to which extent they are required) is missing. This proposal aims at closing this gap. To this end, a generalized metric for the achievable navigation estimation performance shall be derived. On this basis, a theoretical approach shall be developed for the integrated motion planning and navigation, including positions, taking into account the differential constraints of a fixed wing aerial vehicle.