Self-navigating Robot Find its Way on UnfamiliarTerritory by Learning as it Goes

Problem-solving robots making mental maps? We are getting sophisticated.

That's exactly what a team of researchers at Purdue University have created. The result is a robot that can find its way around unfamiliar territory by learning as it goes. The robot will created a two-dimensional map of its new surroundings and compare that map to what the robot has already experienced. If anything looks similar, the robot will ignore that area and move onward.


The technique successfully illustrates the algorithm that the Purdue researchers invented to enable to the robot to problem-solve in the wild. That "wild" isn't all that wild, actually. Such a technique is obviously very successful in a maze environment or an office building—settings that have many similar-looking elements. Outdoors, though, is an entirely different matter. Similar outdoor terrains, like farmland or forests, might be prime experimentation possibilities for the the self-navigating robot; but urban environments such as crowded cities will likely still befuddle such machines. Still, it's another step down the road of progress.




Written by Jeff Strickler
Published in Other Tech