When it comes to part-picking robots, they're typically programmed to grasp a specific object in a specific fashion – such as a car part on an assembly line. Scientists from MIT and Princeton University, however, have developed a system that allows robots to grab random objects from a bin, then identify what the objects are and where they should go.
Known as the Pick-and-Place system, it incorporates a standard robotic arm equipped with a custom gripper and suction cup, along with a set of overhead cameras. The system uses those cameras to look down into a cluttered bin of objects, and then utilizes machine learning algorithms to differentiate those objects from one another, and to ascertain the best way of grasping each one.
Depending on the size and shape of objects, the arm can suction onto them from above or from the side, it can grip them from above, or for items that are flush against one wall of the bin, it can grip them from above while also sliding a flexible spatula between them and the wall.
Once the arm has lifted the object out of the bin, the cameras examine it from various angles. Those images are compared to a database of images of known objects, allowing the system to determine just what it is that the arm is holding. Upon being identified, the object is then moved to an appropriate separate bin.
"This can be applied to warehouse sorting, but also may be used to pick things from your kitchen cabinet or clear debris after an accident," says MIT's Prof. Alberto Rodriguez. "There are many situations where picking technologies could have an impact."
The arm, which is now also equipped with tactile sensors, can be seen in action below.
Want a cleaner, faster loading and ad free reading experience?
Try New Atlas Plus. Learn more