The problematics of precise pointing and more specifically an attitude control is present sincethe first days of flight and Aerospace engineering. The precise attitude control is a matter ofnecessity for a great variety of applications. In the air, planes or unmanned aerial vehicles needto be able to orient precisely. In Space, a telescope or a satellite relies on the attitude control toreach the stars or survey the Earth. The attitude control can be based on various principles, pre-calculated variables, and measurements. It is common to use the gyroscope, Sun/Star/horizonsensors for attitude determination. While those technologies are well established in the indus-try, the rise in a computational power and efficiency in recent years enabled processing of aninfinitely more rich source of information - the vision. In this Thesis, a visual system is used forthe attitude determination and is blended together with a control algorithm to form a VisionBased Attitude Control system.A demonstrator is designed, build and programmed for the purpose of Vision Based AttitudeControl. It is based on the principle of Visual servoing, a method that links image measure-ments to the attitude control, in a form of a set of joint velocities. The intermittent steps arethe image acquisition and processing, feature detection, feature tracking and the computationof joint velocities in a closed loop control scheme. The system is then evaluated in a barrage ofpartial experiments.The results show, that the used detection algorithms, Shi&Tomasi and Harris, performequally well in feature detection and are able to provide a high amount of features for tracking.The pyramidal implementation of the Lucas&Kanade tracking algorithm proves to be a capablemethod for a reliable feature tracking, invariant to rotation and scale change. To further evaluatethe Visual servoing a complete demonstrator is tested. The demonstrator shows the capabilityof Visual Servoing for the purpose of Vision Based Attitude Control. An improvement in thehardware and implementation is recommended and planned to push the system beyond thedemonstrator stage into an applicable system.


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Vision Based Attitude Control


    Beteiligte:

    Erscheinungsdatum :

    2018-01-01


    Medientyp :

    Hochschulschrift


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Klassifikation :

    DDC:    629



    Attitude control system of driverless vehicle based on vision navigation

    Qiu, Heng / Li, Yang / Guo, Chongyu | IEEE | 2020


    Vision-Aided Attitude Control for Space Debris Detection

    Felicetti, Leonard / Emami, M. Reza | AIAA | 2017


    Attitude measurement by artificial vision

    Truchetet, F. / Aubreton, O. / Gorria, P. et al. | Tema Archiv | 2006


    Calculating Attitude from Horizon Vision

    Cornall, T. D. | British Library Conference Proceedings | 2005


    ATTITUDE CONTROL SYSTEM, ATTITUDE CONTROL METHOD, AND ATTITUDE CONTROL PROGRAM

    TAMAKOSHI HIROSHI | Europäisches Patentamt | 2023

    Freier Zugriff