SignScribe: A Wearable Antenna for Sign Language Transcription

This paper presents the design and implementation of a wearable, flexible slotted patch antenna for real-time American Sign Language (ASL) transcription. The antenna operates over a frequency range from 500 MHz to 3 GHz and is fabricated on a Polyethylene Terephthalate (PET) substrate. It is integrated into a glove, with circular copper traces positioned on the fingertips. As the user executes ASL gestures, their hand movements dynamically cover the antenna slots, altering the surface current distribution and, consequently, the reflection coefficient (|S<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sub>|).The antenna is connected to a Bluetooth-enabled vector network analyzer (mini-VNA) that monitors (|S<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sub>|) in real time. The resulting |S<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sub>| data for each gesture are processed using a sequential neural network-based machine learning algorithm, achieving an ASL-to-English translation accuracy of 88.2%. Furthermore, a mobile application interfaces with the system to display the translated words through both images and audio. The complete system was fabricated and tested, validating the effectiveness and feasibility of the proposed ASL transcription method.