Robotic System Zeroes in on Objects Most Relevant for Helping Humans
Using a novel relevance framework developed at MIT, the robot identifies and prioritizes objects in the scene to autonomously assist humans in a seamless, intelligent, and safe manner.
For a robot, the real world is a lot to take in. Making sense of every data point in a scene can take a huge amount of computational effort and time. Using that information to then decide how to best help a human is an even thornier exercise.