Development of a localization system based on ArUco markers for a small space platforms test bench

CubeSats are an increasingly common reality, both in the academic and industrial fields, thanks to their possibility of carrying out a wide variety of missions, with a “low cost” and “fast delivery” approach. The verification and validation phase usually take place after the assembly phase or on the individual subsystems, losing the possible interactions between them. Therefore, arises the need of a platform capable of reproducing simultaneously as many aspects as possible of the CubeSat operating environment and capable of simulating the undeveloped subsystems through a virtual model. This facility would allow to support the development of the satellite throughout its life cycle, further reducing time and cost. CAST (CubeSat Advanced Simulator and Test bench), a project born within the CubeSat Team, aims to improve and boost the verification process of a CubeSat, providing an integrated environment. Thanks to a modular architecture and the in-the-loop verification approach, CAST is composed of four main elements: 1) a Control Centre, to simulate the desired modules; 2) a Main Board System, to manage other ground support equipment and interfacing to the test object; 3) a Sun Simulator, to emulate the sun radiation in different scenarios; 4) a Frictionless Table System, to simulate proximity manoeuvres, thanks to a support structure, the TowerSat. This thesis proposes the design of a localization system for the TowerSat, the LocSys. The TowerSat is free to move in a limited environment, but it is necessary to monitor its position and orientation, both for safety reasons and, possibly, to validate the results elaborated from the test object. LocSys adopts optical cameras, as they do not interfere in any way with the test object. To increase the efficiency and the accuracy of the system, some binary fiducial markers have been adopted, called ArUco markers. Two software have been developed in Python programming language: the first implements an architecture for a single camera, in order to test the performance of a such kind of system; the second software implements a multi-camera architecture, to cover a wider area, overcoming the limits of the field of view of a single device and increasing the LocSys accuracy. Both software implements the possibility to calibrate the optical cameras, through a properly developed calibration tool. The calibration phase is the most crucial, greatly influencing the results. Finally, some tests were carried out to validate the system, obtaining an average error less than 0.5 cm for the position and less than 1 deg on the orientation.