Robot with Raspberry Pi can learn to separate waste with TensorFlow

Developers have created a robot that can learn to separate waste using TensorFlow, the open source library for machine learning. The robot can recognize objects and throw them into one of the three bins.

The robot, little more than a slider, camera and tilting platform, is controlled by a Raspberry Pi and thanks to TensorFlow, the machine can send data to the cloud to improve object recognition. In the setup, the developers show how the robot throws waste into the correct bin. The robot is an initiative of Recycle.ai. The developers are building Rai, a robot that can separate waste by itself. The developers have not yet put the source code online, so unfortunately we cannot yet put a waste-separating robot under our desk.