New robotic hand ‘offers dexterity and learning capabilities’

Computer science experts and engineering researchers have developed a five-fingered robotic hand with advanced dexterity and learning capabilities.

Created by the University of Washington, the hand is not only able to perform dexterous manipulation – including rolling, pivoting, bending and sensing friction – but can also learn from its own experience without needing humans to direct it.

This was made possible by developing an accurate simulation model that enables a computer to analyse movements in real time, with machine learning algorithms helping the hand calculate physics and plan which actions it needs to take to achieve the desired result.

It means overseers no longer need to programme each individual movement of the robot's hand in order to complete a single task.

Lead author Vikash Kumar, a University of Washington doctoral student in computer science and engineering, said: "Hand manipulation is one of the hardest problems that roboticists have to solve. A lot of robots today have pretty capable arms but the hand is as simple as a suction cup or maybe a claw or a gripper."

Though the technology is currently too expensive for routine commercial or industrial use, the research allows engineers to explore theories and test innovative control strategies.