Home > Europe, Space, Technology > Spacecraft to be controlled by artificial intelligence

Spacecraft to be controlled by artificial intelligence

February 14, 2011

It is a concept that had fatal consequences for the astronauts in the science fiction movie 2001: A Space Odyssey after their spaceship’s artificially intelligent computer reasoned it had to kill them in order to continue the mission.

Yet despite this warning from Arthur C Clarke and director Stanley Kubrick, The European Space Agency now hopes to use real-life artificial intelligence to control future spacecraft.

British engineers, supported by ESA, are developing control systems that can be used in satellites, robotic exploration vehicles and spacecraft capable of controlling themselves.

The space vehicles will be able to learn, identify problems, adapt during missions, carry out repairs and take their own decisions about how best to carry out a task.

Details of the research have emerged as ESA prepares to launch the second of its Automated Transport Vehicles to deliver supplies to the International Space Station later this month.

The ATV2, which was designed and built by space company Astrium, will follow a preprogrammed route to the space station before using on-board sensors and collision-avoidance systems to dock safely space station.

ESA has also revealed it has plans to build its first spaceship capable of carrying humans into space and returning them safely back to Earth.

Professor Sandor Veres, an expert in autonomous control systems at Southampton University who has been leading the project to develop artificially intelligent control systems, said the technology could eventually find its way into spacecraft used to transport human crew.

He said: “I think to begin with we are going to see it used in robotic, unmanned missions and in many ways will reduce the need for humans to be in space.

“Communication satellites and space exploration missions that currently need around the clock monitoring by human controllers on Earth will be able to operate of their own accord, which would greatly reduce the cost. We are looking at the next generation of vehicles from the ATV.

“We have been able to demonstrate that the prioritisation and decision making that only humans could make before can now be done by machines in a limited knowledge area.

“So we are not talking about them learning all about physics, but in terms of its movements and the tasks it has to perform, the system can have goals, prioritise, made predictions about the kind of problems it might face and calculate very quickly how to deal with those. It can assess far more information than a human engineer could.”

Professor Veres has been developing the spacecraft control system, called sysbrain, so that it can use natural language to obtain new instructions or new information. It means the spacecraft’s control system would be able to read documents written in English rather than needing specially programmed code to be uploaded.

He added: “The system could even go onto the web, read up to date information and learn from it.

“We have combined this with human like reasoning so that it has perception about the world, its goals and the ability to prioritise those goals before deciding what it wants to do.”

Artificial Intelligence in spacecraft has been a common theme in science fiction movies and books, but the most worrying example was the computer HAL 9000 from the Space Odyssey movies and books.

In 2001: A Space Odyssey, HAL starts to malfunction and kills the crew on the spacecraft before they have a chance to disable the computer.

Professor Veres said any systems used in the real world would be programmed to ensure they did not make decisions that put humans in danger.

Dr Wolfgang Paetsch, head of the ATV production and development at Astrium, said the company was also working on artificial intelligence systems for use in spacecraft.

He said: “Safety is everything in manned space flight. One of the biggest challenges facing artificial intelligence at the moment is ensuring really safe.”

ESA has already tested a form of artificial intelligence during its Mars Express mission to help scientists work out the best time to download data from the spacecraft.

Alessandro Donati, head of advanced mission concepts and technologies at ESA’s Space Operations Centre in Germany, said it hopes to use artificial intelligence software in future planet exploration rovers and other spacecraft.

He said: “When they find a rock that is of particular interest they can decide on their own to take a snapshot and reprogram their other planned activities – without having to wait for new commands to be sent the following day telling them what to do.”

In 2008, the first ATV delivered supplies to the space station after successfully docking with it while orbiting the earth at a speed of 17,400 miles per hour.

Using 28 separate thrusters it was able to manoeuvre itself to safely combine with the metre wide air lock to dock with the space station.

The second ATV, which is also about the same size as a double decker bus and can carry 7 tonnes of cargo, is due to be launched on 15 February and is expected to confirm the technology is safe and reliable.

The current generation of ATV spacecraft are designed to burn up in the atmosphere after they have delivered their cargo to the space station.

But ESA has started work on a similar vehicle that would be capable of bringing cargo and experiments back to Earth by putting them into a re-entry capsule.

Nico Dettman, ESA’s head of the ATV production programme, has confirmed that it is due to propose a version of the ATV that can carry humans and could be one of the spacecraft that will replace Nasa’s shuttle when it is retired.

He said: “”We are expecting to propose something to member states on an automated re-entry vehicle and crew transportation.”

%d bloggers like this: