When This Robot's Leg Gets Broken, It Can Start Walking Again In Minutes
"When animals lose a limb, they learn to hobble remarkably quickly," reads a post on science blog arXiv. "And yet when robots damage a leg, they become completely incapacitated." Two engineers are looking to change that, mimicking the behavior of those animals. Together, Antoine Cully and Jean-Baptiste Mouret have designed a robot that's capable of learning to live with damaged body parts, through a process of "intelligent trial and error."
"This new technique will enable more robust, effective, autonomous robots," said the engineers, who explained how - just a few minutes after having one of its limbs broken - their robot could learn how to walk again.
It's the latest development in a field that's come to be known as adaptive robotics; the design of robots which are capable to analyze, understand, and react to a changing environment. To say that the creation of such robots is a challenging process would be putting it lightly. Traditional robots are logical, designed for very specific functions and made to operate within a very narrow range of situations.
Uncertainty is the bane of their existence.
An adaptive robot - particularly one that could learn to 'live' with structural damage - could open up a whole new world of possibility - though the most obvious application of such technology is military in nature. Imagine, if you would, a robotic soldier that can continue attacking no matter how injured it becomes - only complete dismantlement would bring an end to its assault. You'd be forgiven for feeling a twinge of terror at such a prospect.
Only a twinge, though.
According to robotics expert Doctor Fumiya Iida of the Cambridge Machine Intelligence Laboratory, warfare is only the tip of the iceberg for adaptive robotics. Machines equipped with better learning algorithms could be used in just about every field, from energy to assembly to health care to space exploration.
"There are lots of applications beyond the military," said Iida. "You can think of robots in extreme environments, so not only in warfare, but in space such as robots on the Moon and Mars, and in nuclear power plants. Think of Fukushima, for example, where humans can't go."
Of course, the power of a robot that can learn to cope with an injury sort of pales in comparison to a machine that's capable of self-repair. That technology, too, is well on its way: BAE Systems is looking into drones that can fix damage sustained in flight, and contain on-board 3D printers to make new parts. On top of that, there's a new plastic in development at the University of Illinois that's capable of fixing itself.
So what I'm basically saying is that before much longer, we could start seeing intelligent, adaptive robots capable of repairing whatever damage one might potentially cause them. Hopefully they're friendly.