In their research, given that the system cannot be tested in a real platform, the scientists have constructed a small bank of tests which allow them to simulate a scenario of this kind. They first designed vision algorithms which employ satellite model vertices as characteristic points. Then, they perfected the technique to design a new vision algorithm that uses the information about the texture around the points of interest of the image to associate them with the points in the model. “The latter technique is quite robust against noise, changes in the background and variations in illumination”, Abderrahim noted. Furthermore, he stated, “that the speed of the algorithm allows it to be used in real time”.
The idea behind this research came up thanks to the relation between the UC3M Robotics Lab and the engineers of the company GMV, which is dedicated to the space sector, and the existing need in the market for a system of this kind. "Today there is a growing need to provide satellite maintenance services, without relying on manned missions, which are very costly and involve risks. Robotics is the natural solution”, Professor Abderrahim concluded.