A smart glove infused with wearable electronic components can turn the American Sign Language (ASL) alphabet into readable text. The device wirelessly transmits the translation and displays it on additional electronic devices. All this technology totals for less than $100.

The findings were recently published in the journal PLOS ONE, penned by Darren Lipomi, Timothy O’Connor, and their fellow colleagues from the University of California, San Diego (UCSD).

In order to track hand positions, cameras and optical systems, which use infrared emitters and receivers, are often employed. Although both have their strengths, emitters and receivers are often expensive and lack flexibility, while cameras eat up a lot of power. Gloves donning a wearable sensor system avoid the above constraints for gesture tracking.

Figure 1: Overview of the gesture-decoding glove. (a) Photograph of the glove. (b) Photograph detailing the breakout boards for the MCU, accelerometer, and Bluetooth on the PCB. (c) A circuit diagram of a voltage divider. (d) Schematic drawing of wireless gesture recognition system and the flow of information. Step 1: a gesture was made and the strain sensors transduced the positions of the knuckles into variable resistance values. Step 2: the variable values of resistance were converted into voltages by the voltage dividers. Step 3: the MCU measured the nine voltages and, through a binary comparison process, used them to generate a nine-bit key. Step 4: the binary key was used to determine which letter was to be transmitted wirelessly. (Image and Caption Credit: Timothy O'Connor et al (2017))

The ASL-savvy glove uses nine flexible strain sensors, one on the thumb and two on each finger. The sensors work together to detect knuckle articulation. An onboard microprocessor is in charge of computing the ASL letter for each detected position. The Bluetooth radio then takes control, transmitting the text for display.

The electronic glove accurately identified all 26 letters of the ASL alphabet, working alongside an accelerometer and pressure sensors. Results remained accurate after the knuckles were bent a maximum of 1,000 times before sensor fatigue took place.

Adding to the glove’s features, it also was able to move a virtual hand. When a user wearing the device signed “UCSD,” a virtual hand mimicked the motions accurately. “This application suggests new ways in which stretchable and wearable electronics can enable humans to interface with virtual environments,” according to the research.

In addition to the ASL translations, the glove may also serve as a test system for assessing the attributes of new materials and stretchable hybrid electronics.