Robots are surpassing people in numerous areas, but the human hand is still superior. Robots have trouble picking up irregular objects. Now, a new grasper designed at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) could give robots an edge. The origami “magic ball” gripper doesn’t try to imitate our hands, but it can still hold 100 times its weight, Extreme Tech reported.
The origami gripper is an application of what’s commonly called “soft robotics.” As the name implies, these machines use flexible materials like rubber instead of cold, unyielding metal. The new CSAIL contraption looks like a floppy paper flower, but it contracts to grab objects with incredible strength. It doesn’t matter what these objects are as long as you can wedge part of them inside the cone.
The gripper consists of three parts: a folding internal skeleton, an exoskeleton of rubber, and the flexible connector. It works by pumping gas into and out of the airtight interior space. This simple construction means the team can quickly and efficiently iterate the design to improve functionality in various settings.
Companies like Amazon are incredibly interested in developing robots that can pick up irregular objects, and soft robotics is probably the best way to do it. Currently, the team says the cone-shaped gripper can hold almost any object with 70 percent or less of its diameter and up to 100 times its weight. It’s also much gentler than a traditional robot arm, able to lift objects that are both heavy and delicate like glass bottles and food. CSAIL tested the magic ball gripper with various objects like bottles of water, fruit, aluminum cans, and so on. This robot could almost unpack your groceries without scrambling the eggs.
The magic ball gripper is currently best with cylindrical objects like bottles and cans, but it can successfully lift irregular objects so long as part of it fits in the cone opening. Larger flat objects like books are tough, though. Humans are also operating the gripper in the demo above, which is necessary because many of the objects tested are too large in certain dimensions. The team plans to add computer vision to the grasper, which would allow it to find the correct angle at which to grab an object.
