When two vehicles developed by a Canadian robotics firm arrive at a narrow door at the same time, they have a friendly way to decide who should pass through first.


The key is to use a team of robots with an elected leader who makes decisions that are best for the group as a whole. Representatives from the Ottawa-based company Frontline Robotics, who will present their polite robots at the RoboBusiness robotics conference in Cambridge, Massachusetts, this May, say their software should be ideal for a variety of military and civilian applications.



“Wherever the task is dull, dangerous and dirty, robots should be in the front line,” says Frontline’s head, Richard Lepack. He sees the company’s robots helping to investigate scenes of terrorist attack or nuclear accidents, or aiding deep-sea construction.



Many researchers are seeking to develop robots that work in packs, coordinating their actions in response to one another. This way of working, called distributed intelligence, resembles the way that social insects such as ants or bees collaborate in foraging or nest building.



It has been shown that such groups can evolve problem-solving strategies that are more efficient than those in which each robot acts independently.



But the robots in such groups usually don’t make decisions; they simply respond in a programmed way to a set of input signals. The collaboration between Frontline’s robots is different: two robots facing a doorway have to be slightly more flexible.



More here.

0