Making robots that adapt like animals

David Taft for redOrbit.com – Your Universe Online

The main difference between a living creature and a machine is their reaction to external stimuli. While the tissues that make up a biological organism are constantly reacting to changes in the environment on a cellular and even molecular level, the composite or metal structure of a machine remains inert.

However, scientists are currently working to bridge the gap between artificial and biological physiology.

Antoine Cully and Jean-Baptiste Mouret, from the Pierre and Marie Curie University in France, in collaboration with Jeff Clune from the University of Wyoming and Danesh Tarapore also from the Pierre and Marie Curie University, have developed some promising solutions. Their new paper in the scientific journal Nature, entitled “Robots That Can Adapt Like Animals,” documents the capability for autonomous robots to recover from major damage in under two minutes.

Mouret stated that, “When injured, animals do not start learning from scratch. Instead, they have intuitions about different ways to behave. These intuitions allow them to intelligently select a few, different behaviors to try out and, after these tests, they choose one that works in spite of the injury. We made robots that can do the same.”

Ruling out behaviors

The new T-Resilience (transferrable resistance) algorithm allows the robot to self-model its own system and take damage into account while developing new patterns of movement to compensate for the injury. Using this software, a robot with a damaged limb will calculate the efficiency of several different gaits in order to move at optimal speed.

Cully described how, “Each behavior it tries is like an experiment and, if one behavior doesn’t work, the robot is smart enough to rule out that entire type of behavior and try a new type. For example, if walking, mostly on its hind legs, does not work well, it will next try walking mostly on its front legs. What’s surprising is how quickly it can learn a new way to walk. It’s amazing to watch a robot go from crippled and flailing around to efficiently limping away in about two minutes.”

This new set of behaviors will allow autonomous robots and even common machinery of tomorrow to face harsher environments while still carrying out their appointed tasks. Algorithms like the T-Resilience program may have a huge effect on the way manufacturing, military and rescue robots operate in the future.

—–

Follow redOrbit on Twitter, Facebook, Google+, Instagram and Pinterest.

redOrbit.com
offers Science, Space, Technology, Health news, videos, images and
reference information. For the latest science news, space news,
technology news, health news visit redOrbit.com frequently. Learn
something new every day.”