Computer Vision Technology Used to Capture Satellites

Scientists at the Universidad Carlos III in Madrid have developed a new system for docking and capturing space satellites that is based on robotics and computer vision technology. The technology is able to autonomously guide a space vehicle to dock and capture the satellites.

Within the framework of this research project, the scientists have developed algorithms and strategies to dock and capture a disabled satellite through the use of a space vehicle, called a "chaser," which autonomously carries out this function.

"In this way," said Mohamed Abderrahim, the leader of the research project, "we attempt to carry out all the phases of a satellite inspection and maintenance mission with another autonomous vehicle equipped with a robotic system for capture and control."

What is most interesting about the designed prototype is its algorithmic part, according to the researchers, as they are not involved with the manufacture or launching of satellites. Their true contribution lies in the set of well-defined and ordered instructions that allow satellites to be programmed for the future so they can service others that have become disabled or that allow them to prolong the satellite's working life. These navigation algorithms, based on vision techniques, are capable of indentifying the objects of the search among various objects present, estimating their position and orientation.

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. "The speed of the algorithm allows it to be used in real time."

"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," Abderrahim said.