UCLA bioengineering students Priya Srinivasan and Navid Azodi won the $10,000 Lemelson-MIT Student Prize for developing gloves that translate American Sign Language (ASL) into audible speech in real-time.
The SignAloud gloves use sensors to detect hand movements and finger positioning. The data transmits wirelessly to a computer that processes the gestures through machine learning algorithms, converting them into spoken words and phrases.
Technical Implementation
Each glove contains multiple sensors that capture motion data at high frequency. The system recognizes individual letters and common words in ASL, with accuracy rates reaching 95% for tested vocabulary sets.
The wireless transmission occurs via Bluetooth to connected devices running the translation software. Processing happens in under one second from gesture to speech output.
Market Potential and Next Steps
Over 500,000 Americans use ASL as their primary language, creating a substantial market for translation technology. Current professional interpreters charge $150-300 per session, making automated solutions cost-effective for regular communication needs.
Srinivasan and Azodi plan to expand the vocabulary database and reduce hardware costs through mass production. They aim to launch consumer testing within 18 months while pursuing additional patent protections for their sensor configuration.