North Carolina State University and University of Utah researchers are applying the principles of human evolution to the systems that control robots.
The researchers trained neural networks to play Capture the Flag (outdoor game often played by children where two teams have a flag each and the objective is to capture the other team’s flag, located at the team’s “base”, and bring it back to their own base). Once the networks had mastered the game, they moved them to the robots’ computers. These teams of mobile robots, called EvBots (for Evolution Robots), were then able to play the game.
The aim is to develop environment-aware autonomous robots able to cleare minefields or find heat sources in a collapsed building within 3 to 6 years. But the researchers want to build controllers for robots that adapt to completely unknown environments. And this will not happen before 10 or maybe 50 years.