Tech Students Build Robotic Hand as Sign Language Learning Tool

Damien Gaudry never took a course in sign language, but that didn't stop him from wanting to build an instructional tool that could lend a hand to children who want to learn.

Gaudry and two other Georgia Tech graduate students - Cindy Perreira and Russell Marzette - built an 18-inch tall robotic hand that could eventually sign the 26 letters of the English alphabet.

"We wanted to develop an interactive and fun device to teach children sign language, whether they are hearing impaired or not," Gaudry said. "We thought it would help them to visually see the signs rather than having to look at them in books."

The students worked on the project for two months this year as their final project in a graduate-level introductory course on "mechatronics" - an emerging technique in engineering that deals with the study of integration of familiar mechanical systems and components with new electronic components and intelligence-based software.

Students enrolled in the course, offered in Georgia Tech's School of Mechanical Engineering, must design, build and present a mechatronic innovation that fuses mechanics, electronics and information technology. In the class, students are taught to engineer for complete "smart product" development for the market.

Although the robotic hand project has a long way to go to market, the students have accomplished much so far. During their project presentation to the class in April, the students explained they were able to program the hand to make nine different symbols: the letters "A" to "D," and numbers one to five.

That's no small feat, considering the engineering involved in building a giant hand that signs.

"I think the most difficult part was integrating electrical components and computer-based systems into one mechanical system," said Marzette, one of the three members on the team.

"Getting the electronics to work together properly was a challenge," said Gaudry, who hopes the team can find time to take the project further. "Fifteen stepper motors are built in to control the hand and each motor has its own circuitry."

With fingers made from soldered brass tubing and rods, and strung together by cables connected to a series of small motors, the robotic hand is surprisingly agile. It takes about five seconds to sign a letter.

The hand rests on a microprocessor unit that controls the motion and all the circuitry to power the motors. The unit includes an LCD panel and a series of switches that allows the user to enter the letter or number they would like the hand to display.

Gaudry scrolls through the letters and numbers and selects the letter "C." Once selected, the motors begin to wind up the cables in the fingers, bending the joints until the hand produces the symbol, similar to the way a mechanical puppet might be manipulated. After the letter is signed, rubber bands connected to the back of each finger return the hand to the starting open position.

The team members said they entered the class knowing very little about electronics and microprocessor control, but learned more than they imagined. "The class is time consuming, since you have to put in a lot of your own time outside of the lecture portion," Gaudry said.

The robotic hand is just one interesting invention to come out of the mechatronics courses. Other notable inventions have been the self-playing guitar that actually takes requests, a machine similar to an ATM that dispenses money called the "CA$H-o-MATIC," and the automatic card dealer, a machine that deals cards and plays along with human players.

"The mechatronic engineering techniques used in this project are representative of the approach that's used today to design devices and systems that possess a degree of computer-based intelligence," said Georgia Tech Professor Charles Ume, who teaches undergraduate and graduate courses in mechatronics. "We expect to see the integration of mechanics, electronics and computing - mechatronics engineering - increase rapidly in the near future due to consumer demands for smart products and intelligent machines."

Georgia Tech's state-of-the-art Mechatronics Laboratory, which supports these classes, was built with a grant form the U.S. National Science Foundation, a matching fund from Georgia Tech, and financial and equipment support from Siemens.