Eye contact and face engagement are the two most important elements that plays major role in interaction among humans. And this main thing was missing when interacting with robots but now bots too would react according to the given set of instruction(s) that they’d decipher as per the eye and face recognition.
Researchers at IROS made it a reality. They showed how an individual robot obeyed – within a group of bots – an instruction. This breakthrough went ahead in controlling bots in a more natural way.
The bots were implanted with cameras sodden with face detection, which help them determine which human (face) is watching them. Once the face detection is done, the bot looks for the verbal commands that are being issued and the task is performed.
Face recognition is done with a “face score” that the bot calculates given the angle of human looking at them. For instance, in the below given picture, “face score” is given by three bots placed at three different positions in front of the person. Highest face score means the user is staring the bot which further will carry out the given commands. And the same principle goes for gestures as well.
The process works well with selecting multiple bots one at a time or single bots irrespective in a crowd or otherwise.
This approach will help in creating an intuitive command system, which will not require any form of training for a user. Researchers envision that in future, robots will be placed under different team names and naming alone would be enough to give directions to the machines, as is the case with human community.
This sounds really cool and of course futuristic too. Soon we would be hearing naming conventions among robots and the best part will be they would actually be responding to it. Just awesome!