Inspiration
Traditional mobility aids for the visually impaired, like the white cane, are limited. Audio-based tech often causes "auditory saturation," blocking vital environmental sounds. We wanted to create a hands-free, ears-free solution.
What it does
It is a wearable haptic vest equipped with a matrix of 9 vibration motors. It uses a mobile app powered by computer vision to detect obstacles and map the environment, translating this data into real-time spatial vibrations on the user's torso.
How we built it
We developed the mobile application using Flutter and Dart. The real-time object detection is handled locally via TensorFlow Lite (SSD MobileNet V1), while the Google Gemini API provides advanced contextual awareness. The hardware consists of a microcontroller connected to the app, controlling the 9-motor vibration matrix with zero latency.
Challenges we ran into
Synchronizing the hardware with the computer vision model to achieve immediate, reliable haptic feedback without draining the smartphone's battery.
What's next
Refining the hardware prototype, improving the machine learning model's accuracy, and conducting real-world testing.
Log in or sign up for Devpost to join the conversation.