The further we get from Earth, the more space probes and robots will have to think for themselves. But developing autonomous robots is hard without artificial intelligence.
Picture the scene: heaps of sand, gravel and large rocks surround a space probe that looks like it's just landed on a foreign planet and is about to release a robot to explore the terrain.
But this planet isn't real.
It was set up by the German Aeronautics and Space Research Center (DLR) at a multi-purpose facility near Bonn, where ten teams from German universities and industry have just competed in the DLR SpaceBot Cup. Basically, it was a chance for them to show off their latest, high-tech robots.
The planet may be fake, but the concept behind the competition is real enough. And important for further space exploration.
NASA has sent several generations of rovers to Mars. But it takes a long time to send signals from Earth up to the robots on the Red Planet: 15 minutes one way. This means the wait for an answer is at least half an hour long.
If ever we are able to send robots deeper into space, communication will be even more complicated. So scientists want future space robots to be able to work on their own - using artificial intelligence.
The search for water
As with a real Mars rover, the robots at the SpaceBot Cup are looking for water.
The one exception being that in this setting the water is in a blue drinking glass rather than hidden deep inside a rock. The robots have to find the glass, grab it and move it to another location.
Lauron, a six-legged contraption developed at the FZI Research Center for Information Technology at the Karlsruhe Institute of Technology, is especially good at walking.
The developers took inspiration from the Indian stick insect.
"The legs are very flexible and we equipped the front one with a grasper," team spokesman Lars Pfotzer explains.
Lauron won't be reduced to walking on its remaining five legs once it grabs the glass of water, though. It's got a bracket under its midsection, where the robot can store cargo and continue walking on six legs.
For orientation purposes, it uses a laser scanner. The rotating device is located on top of the robot and creates a three dimensional image of the environment.
Figuring out the right steps
Lauron also has a second set of eyes, similar to the camera in the Kinect gaming console.
It sees its environment as a cloud of points or pixels and uses these points to calculate three dimensional images. The robot uses these 3D images to figure out its position in relation to the glass of water and this helps it recognize the glass as an important object.
Once the robot has recognized the glass, it can figure out what it needs to do next: that is, grab the glass and transport it to its goal.
The NimbRo Centauro did just that.
The six-wheeled robot was constructed by a team at the University of Bonn.
NimbRo Centauro located the glass and secured it in a plastic holding fixture complete with a lid.
The audience at the SpaceBot Cup went wild, and the competition announced NimbRo Centauro "hasn't spilled a single drop of water!"
Not just for space
Andreas Birk of Jacobs University in Bremen has used similar robots in danger zones after natural disasters.
His team is specialized in using robots to develop high quality maps of unknown regions - for example, after the Fukushima nuclear meltdown in Japan.
"We've developed a software to create 2D maps, using drones," says Birk. "And we have a cooperation with an American partner that used the software to draw up small maps relatively quickly after the catastrophe."
But the presented a land-based robot at the SpaceBot Cup.
"We're able to show that the robot's gripper can handle objects quite well and is able to keep it in a horizontal position," Birk says.
In fact, most of the autonomous robots on show did well - once they had found the glass - because finding the glass was often harder than handling it.
"It's often the things we think are the easiest things, such moving around and finding objects," says Birk, "which end up being the most challenging!"