Robot engineers at the Massachusetts Institute of Technology (MIT) have found a way to give their new robot a better grip. It’s more adjustable and intuitive than that of previous robots.
The progress was made possible thanks to a model that the researchers developed. It predicts the force with which a robotic gripper has to push on various fixtures in its environment so that it can adjust the grasp it has on an object.
For instance, let’s say that a robotic gripper wants to pick up a pencil by grabbing it at its midpoint, but somehow ends up grabbing it by the eraser end instead. In such a scenario the new robot would uses its environment in order to adjust its grasp.
Rather than follow a more traditional path by putting down the pencil and trying to pick it up again by its midpoint, the newly developed model allows the machine to carefully loosen its grip, then slowly push the pencil on a nearby wall until the robotic gripper slides closer to the desired gripping point.
To enable the machine to predict how an object would move when pushed against a fixed surface, the robot engineers made sure that their model takes into considerations a few different factors, the main ones being the frictional force between the object and the gripper, and the frictional force between the object and the environment.
Other factors worth mentioning are the object’s shape, mass and inertia.
The research team, led by Alberto Rodriguez, assistant professor of mechanical engineering, and Nikhil Chavan-Dafle, graduate student, has dubbed this new approach “extrinsic dexterity”, the main implication being that the technology is a long way from matching the intrinsic dexterity of the human hand.
When a person wants to adjust their grip on a pencil, they can simply let their fingers wonder on the object until they reach the desired gripping point.
The MIT team hopes that their model will be put to good use and allow robots used in the fields of medicine, disaster response and manufacturing to perform more complex tasks for a fairly cheap price.
Rodriguez gave a statement saying that if a person or a company can’t afford to buy a complex, USD 100.000 hand, this model is the next best thing as it “brings some dexterity to very simple grippers”.
Image Source: frankmoss.com
