At last week’s International Conference on Robotics and Automation, researchers at Carnegie Mellon University (CMU) presented work on an object-recognition system that lets a robot sort through a pile of recyclables by hand. Being able to pick out items from a cluttered, disordered environment is no easy feat. And while other robots are now dexterous enough to grasp an egg without breaking it, and some can pick up unfamiliar objects, these systems generally work only if the object in question has been positioned carefully.
The CMU researchers and collaborators from Intel Research Pittsburgh, developed the system, which merges information from several images in order to create a 3D model. Then, by focusing on features like corners or textured areas, an object-recognition algorithm can spot a particular object within a pile of clutter.
When it finds enough matches between features, the algorithm considers the object identified. By looking for features of an object rather than the object as a whole, the vision system can recognize objects faster than many existing systems, says Alvaro Collet Romea, a master’s student at CMU who led the research. It can even identify and pick up objects that are partially obscured. The video below shows a robot called HERB using the system to separate recyclables.
When designing an embedded system choosing which tools to use often comes down to building a custom solution or buying off-the-shelf tools.