A human baby very fast learns to grasp different objects in unstructured environments. Until now this has been a complicated and very slow task for robots. Lerrel Pinto and Abhinav Gupta at Carnegie Mellon University in Pittsburgh has taken up this challenge and invented the Baxter robot. The general purpose was to create a robot that interacts with everyday objects in an intelligent way. In order to make this happen Baxter is equipped with powerful deep learning capabilities, while its body is similar to a robot on the factory floor. “Over 700 hours, Baxter performed some 50 000 grasps on 150 different objects, each time learning whether the approach was successful or not.” From this training Baxter now has reached the level of 80% accurate grasps. More complicated tasks in a common sense approach the robot had lower accurate grasps. Humans still have a great advantage within this field, but one can wonder for how long? Source: MIT Technology Review
Latest News: Baxter Robot teach itself to Grasp objects
Bookmark the permalink.