Nothing says it’s nearly Christmas quite like a 3D-printed robot hand busting out a festive classic on the piano. Fortunately, this year the tradition continues (or, let’s be honest, starts) courtesy of researchers from the U.K.’s University of Cambridge. Using a 3D printer, they have created a soft robot hand — attached to a robot arm — that’s able to play the piano in a variety of styles, including clipped “staccato” and smooth “legato” notes. And what better way to showcase it than with a rendition of Jingle Bells?
While this isn’t the first piano-playing robot we’ve covered, what makes this robot neat is that it can achieve this fairly complex action despite being pretty limited in its movement. For instance, the hand is unable to move its fingers independently. That’s in stark contrast to the real human hand, which is capable of incredibly fine-grained motions thanks to its various points of articulation. The fact that it is therefore able to carry out actions as complicated as playing musical phrases on the piano is a testament to what can be achieved through some enterprising design. When it comes to this robot hand and its attached robot arm, the movement really is all in the wrist.
“Creating robots which can manipulate objects as well as humans can is really challenging,” Josie Hughes, a researcher in Cambridge’s department of engineering, told Digital Trends. “We have explored how the physical structure of the hands — the bones and soft ligaments — provides some intelligence which enables this. By developing methods of 3D-printing a hand structure with soft and rigid components, we can then just actuate the robot hand, such that we can explore the passive behaviors of the hand as it interacts with the environment. This is a far more scalable approach than other approaches to developing complex robot hands, where many motors and really complex control is needed.”
The 3D-printed hand is just the latest in a growing number of “soft robotics,” which differ in material from the more rigid robots we might more typically associate with the word.
“We now want to take this work further to investigate how we can exploit this mechanical complexity of the hand, and how we can use this to enable robots to perform highly complex tasks which robots currently can’t perform,” Hughes said “To advance the technology further we also want to include tactile sensing into the hand — the human hand has an incredible sense of touch — and we want to include this in a robot hand.”
A paper describing the work was recently published in the journal Science Robotics.
- Why teaching robots to play hide-and-seek could be the key to next-gen A.I.
- This quirky humanoid robot can be teleoperated using a VR headset
- Boston Dynamics’ Spot robot shows off its grabbing skills in new video
- The best robots at CES 2021
- Facebook’s new image-recognition A.I. is trained on 1 billion Instagram photos