Teaching robots how to differentiate between different types of garbage is not always straightforward. This video from Dimitrios Kanoulas shows how a mobile manipulator with deep learning and whole body control can collect and sort garbage. It uses a deep neural network (GarbageNet) to detect different recyclable types of garbage.
Researchers used “a grasp localization method to identify the grasp poses of garbage that need to be collected from the ground.” The above video shows it in action.
[HT]