Students Create Robot Hands to Teach Sign Language

From The Post & Courier…

“SC cadet students build robot hands to teach American Sign Language

Mohamed Baghdady leans to the microphone and says the word “one.” A lone mechanical finger on the table responds with a curl, then rises upright.

At another table in The Citadel engineering lab, a complete robotic hand spells the sign language letters for “c-a-t” from instructions typed on a computer. The movement is driven by motors and fishing line.

The hands would work with voice recognition software, replicating the spelling and gestures human hands use to communicate with someone who is deaf.

The work in professor Robert Rabb’s class could lead to another breakthrough in the uses of robotic limbs — a world of prosthetics and  “microbots” diagnosing illnesses inside the body.
But maybe the coolest feature of the hands is you could build them at home. The class final product won’t be a pair of hands; it’ll be an online workshop on the website Instructables’ how-to guide. 
Baghdady, a cadet, is building a pair of hands to communicate as a teaching aid and potentially as a human substitute when an American Sign Language translator can’t be brought in.

“Even a middle-schooler could build this project,” said cadet Paul Vargas. “It’s not only mobile, it’s cheaper to use, cheaper to produce than most robotics, and it fills a need.”
There are a few obstacles to overcome.

The class is now trying to flex the hand wrists — a necessary component to communicating in sign. In early prototypes, moving the hand took so much electricity it left the finger joints unable to return upright, said cadet Zachery Danis.

But the biggest hurdle is a little-realized subtlety in ASL itself.
The “grammar” of sign language includes physical or facial expressions of the interpreter, said Jason Hurdich, an ASL interpreter who gained fame when he interpreted for Gov. Nikki Haley in 2016 as she pleaded for residents to evacuate ahead of Hurricane Matthew.
The first control board the cadets built blew up on them: It needed a shield. The vocal-recognition computer program struggles with words that begin or end with a vowel…”

Read the full article at its source: