Multirotor attitude determination using vector measurements taken from a single camera

• Grant number: 15/04393-2
• Support Opportunities: Regular Research Grants
• Start date: June 01, 2015
• End date: May 31, 2017
• Field of knowledge: Engineering - Aerospace Engineering - Aerospace Systems
• Principal Investigator: Davi Antônio dos Santos
• Grantee: Davi Antônio dos Santos
• Host Institution: Divisão de Engenharia Mecânica (IEM). Instituto Tecnológico de Aeronáutica (ITA). São José dos Campos , SP, Brazil
• City of the host institution: São José dos Campos
• Associated researchers: Carlos Henrique Quartucci Forster ; Helio Koiti Kuga

Abstract

This project will deal with the attitude determination (AD) of multirotor unmanned aerial vehicles (UAVs) using vector measurements taken from a single camera. The camera is fixed to the vehicle structure and points downwards. The AD from vector measurements consists in a well-investigated topic in the literature related to aerospace systems. In this area, the typical vector measurements are from the direction of the Sun, the direction of nadir, the direction of distant stars, and the geomagnetic field density. The AD methods currently applied to multirotors are, in general, based on, at most, two vector measurements: the direction of gravity (using accelerometers) and the direction of the geomagnetic field (using magnetometers). We argue in the text that both measurements show deficiencies which limits the performance and the reliability of systems based on those methods. The present work proposes alternatively a framework for attitude determiantion of multirotors from vector measurements obtained a single strapdown camera, pointing downwards. Such vector measurements correspond to the direction between the vehicle's center of mass (CM) and landmarks. The methods derived from the aforementioned framework will be evaluated in two parts. In the first part, computational simulations in MATLAB/Simulink and Simulink 3D Animation will be carried out. In the second part, the method will be evaluated on experimental texts of a real-time implementation of the system on a Raspberry Pi computer. (AU)