Inspiration
The idea for AccessMate was born from observing the daily challenges faced by individuals with visual and mobility impairments
What it does
Whether it's navigating a crowded train station or trying to locate a restroom in a public mall, these situations often lead to frustration, dependency, or even complete avoidance of public spaces.
How we built it
Python – For backend logic and object detection
OpenCV – For real-time image recognition
TensorFlow Lite – Lightweight ML models for edge devices
Google Maps API – For route guidance and geolocation
Arduino + Ultrasonic Sensors – For physical obstacle detection
Text-to-Speech (TTS) – For vocal instructions
React Native – To build a cross-platform mobile app
Challenges we ran into
Sensor Noise and Accuracy: Getting reliable readings from ultrasonic sensors in noisy environments was a major challenge.
Latency in Object Detection: Ensuring near real-time feedback without draining mobile battery.
Voice Recognition Errors: In noisy places, speech inputs often failed, leading us to add keyword-based controls.
Inclusive UX Design: Making an interface that’s fully operable through voice, vibration, or screen-readers required several iterations.
Log in or sign up for Devpost to join the conversation.