Digital language translators are used every day. However, there is one dialect that still eludes the automation.
Engineering students, Ranjay Krishna, Seonwoo Lee and Si Ping Wang, from Cornell University are attempting to develop a very different type of translator, one that helps those who cannot make use of auditory signals. The students have created a sign language translator that converts hand gestures to their corresponding letter symbol and sound.
This translator is in the form of a glove for the hand and circuitry that fasten to the forearm. The glove itself contains nine flex sensors, four contact sensors, one x-y axis accelerometers and one z-axis accelerometer. The flex sensors are positioned around the upper/lower knuckles and the contact sensors at the tips of the fingers to distinguish between gestures. The accelerometers are needed because some letters vary only on movements of the hands and other letters vary only on the orientation of the hand.
An ATMega644 Microcontroller is used to analyze the signal from the electromechanical sensors and send transmission requests to the transmitter. The device is made wireless using a Radiotronix transceiver. All of this makes up what they call the Detection Unit, and it is simply strapped to the forearm. The signal is then received by the base station with outputs the results to an LCD screen as well as transmits the signal to a computer via USB where it can also be outputted as sound through the computer speakers. The students used Matlab, Java and C for all their coding.
As far as can be seen, this device only converts gestures to letters so more development is needed for those gestures that represent nouns or actions but no doubt this is a start towards a world where we can all communicate a little easier.
Cabe
