Menu
38,000 kilometres above Earth — beyond the Hubble telescope, beyond the international space station, and beyond even the farthest-flung geostationary satellites — is a graveyard. Littered with over 3000 defunct satellites, this is where geostationary satellites come to die. These retired behemoths invested their final three months’ of operational fuel into reaching this orbit, just 300 kilometers above the geostationary ring. They will remain here potentially for millions of years, silently circling our great-great grandchildren’s world — the loneliest trash in the solar system.
Satellites, like all hardware, inevitably break down. They run out of fuel, experience a malfunction, or become obsolete. After millions — or billions — invested, even relatively minor malfunctions can render entire satellites inoperational, forcing their owners to make difficult choices (if, indeed, any options remain) about how and where to abandon their suddenly worthless hardware. But these tough choices might soon, themselves, be obsolete. That’s thanks, in part, to SnT’s Computer Vision, Imaging & Machine Intelligence research group who have partnered with space-industry start-up LMO to develop computer-vision technology that could help enable automated satellite repair missions.
“We envision a future where we can offer satellite operators an alternative to early retirement when things go wrong.”
“We want our spacecraft to be able to execute in-orbit satellite repairs and other services, like re-fueling or re-orbiting.” The ability to repair and resuscitate dying satellites would be an important, and likely disruptive, change to an industry that has until now regularly accepted the total loss of their investments.
And a future where satellites can be repaired is a future where satellite missions can be flexible. With the introduction of standardised parts, satellites could even be constructed or re-configured in orbit. This could expand the realm of the feasible — enabling new applications for satellite technology and innovative new-space business models to bring the industry into the next century.
But to get there, the robots that conduct the repairs need to see. Astronauts, like those on the Challenger spacecrafts in the 1980s, used to occasionally conduct in-person satellite repairs. Today, our eyes in the sky are limited to sensors — and that is where computer-vision expert Prof. Djamila Aouada and her team come in.
“This is a very new field and there is a lot of work to be done to grow the role of computer vision in the space sector.”
“Our work is typically applied to security solutions on the ground, so bringing our research into space will give us many new research opportunities,” said Prof. Aouada, principal investigator on the project. Computer vision technologies will allow space servicing vehicles to not just see the satellites they work on, but also to make sense of that visual data. At the end of the day, the goal is for these vehicles to be able to independently — and automatically — assess a satellite and make any perform any repairs or maintenance. Through this research collaboration, Prof. Aouada and her team will develop the foundational technology and LMO will put this exciting new capability in the hands of spacecraft manufacturers.
