Datenbestand vom 15. November 2024

Warenkorb Datenschutzhinweis Dissertationsdruck Dissertationsverlag Institutsreihen     Preisrechner

aktualisiert am 15. November 2024

ISBN 9783843946261

72,00 € inkl. MwSt, zzgl. Versand


978-3-8439-4626-1, Reihe Luftfahrt

Franziska Funk
Passive Cloud Detection for High Altitude Pseudo-Satellites

170 Seiten, Dissertation Universität der Bundeswehr München (2020), Softcover, A5

Zusammenfassung / Abstract

While in a conventional aircraft the pilots’ human sensory inventory enables situational awareness and decision making and is already supported by technical sensory means, such functionality becomes inevitable when it comes to unmanned flight. Respective activities are covered by research and development in the field of Sense & Avoid, which technologies refer typically to other air space users or, for low altitude flight, ground hazards. In this work however, the interest is on certain natural weather phenomena, more specifically clouds.

This thesis proposes a visual sensor-based detection system specifically to be used on-board of so-called High Altitude Pseudo-Satellites (HAPSs). HAPS is an extremely light-weight platform, which, being solar powered, is built for long-term missions in high altitudes up to the stratosphere. For this platform the weather phenomenon of clouds is of high importance, as their presence directly affects operation. The proposed system utilizes two mapping modes for both largescale cloud awareness and three-dimensional localization of each cloud object in the vicinity.

The design of this system and its functional use was based on both platform limitations and HAPS specific mission characteristics, which largely separate this application from typical unmanned aerial vehicles. Following the design, this work describes the implementation of these methods. In order to validate the system and approach, a concept for validation was developed and test environments designed. Subsequent analyses and results of the designated test cases showed, that various research gaps were successfully closed. Tests in both simulation and hardware-in-the-loop hereby compare performances for critical methods, such as cloud segmentation and distance estimation, and their robustness was shown. Finally, full functionality of both modes of operation was successfully demonstrated.