Last month an international research group reported its progress on a new technology which might forever change the way we explore the surfaces of extraterrestrial planets. As we are focusing on Mars (and consequently the Spirit, Opportunity and Curiosity rovers) in class, I figured this would be the perfect topic for Wednesday’s post.
Headed by Patrick McGuire of Berlin, this team has been investigating autonomous modes of robotic research in an attempt to make extraterrestrial exploration more efficient. As it is now, thanks to the distance between us and other planets in our solar system, as well as the speed of modern means of communication, there is a severe lag effect in rover operations. For example, it takes about 14 minutes for a simple movement command sent from Earth to reach a rover on the surface of Mars. As you can imagine, this makes the exploration of the Martian surface an incredibly slow and tedious process.
In an effort to speed things up, McGuire and company have been working on a new technology, referred to as “Autonomous Computer Vision”, that enables a computer to recognize oddities in the patterns of a landscape. The hope is that eventually they will be able to program a rover to pick out strange geological features on the surface of other planets without having to refer back to Earth for human confirmation, which at the moment is a half-an-hour ordeal.
As of last month, working in an abandoned, “Mars-like” mine site and using only a laptop and a cellphone camera, McGuire’s team was able to write a program that, by differentiating color shading and examining pixel redundancy, could classify images that a human had identified as either “normal” or “unique” with 90% accuracy. Using only a laptop and a cellphone camera.
The term “Cyborg Astrobiologist” comes from the fact that McGuire’s team essentially “trained” the computer to recognize shapes. They would take a series of pictures and then label each as normal or unique, and submit the classifications to the computer, which would store them for memory. As they processed more and more images, the computer gradually became more refined in its definitions of normality. Thus, the technology can be called the combination of mechanical and biological systems, hence the “cyborg”.
The greater implications of such technology are perhaps even more interesting than its practical applications to space rovers. The fact that even prototype technology exists that enables a computer to, in all practical sense of the word, see, makes you step back and wonder how long it will be until androids have conquered the globe. Well not really, but you can see the connection. Research that will enable a robot to operate autonomously in response to purely optical data is a significant step towards a self-sustaining, functional platform. Now all it needs is an artificial brain and… bam. iRobot. Just something to think about.
http://www.insidescience.org/content/cyborg-astrobiologist-study-alien-planets/1419
