The University of Maryland Researchers Introduce a Novel Method, Called TERP, for Reliable Robot Navigation in Uneven Outdoor Terrains Using Deep Reinforcement Learning (DRL)

Autonomous mobile robots are being used more and more in the real world for things like indoor and outdoor surveillance, search and rescue, exploring planets and space, extensive agricultural surveys, etc. For each of these uses, the robot needs to be able to work on different types of terrain, which can be described by things like color and texture and things like changes in elevation, slope, etc.

The unevenness and slope of the ground mainly determine a robot’s stability, which …

learning machinelearningnews navigation reinforcement reinforcement learning researchers robot university university of maryland