Assistive AI and IoT · Grades 10 to 11
Whistle to wrist
A wristband that turns a referee's whistle into a vibration, so a hearing-impaired footballer gets the call at the same moment as everyone else.
- prize, Young India Innovates 2026
- 1st
- prize, Young India Innovates 2026
- whistle detection accuracy
- ~90%
- whistle detection accuracy
- latency from whistle to wrist
- ~0.5 s
- latency from whistle to wrist
01 The person
Why it exists.
On a turf with the kids I coach, I watched a player keep running after the whistle. The system listens on a microphone, recognises the whistle with a pre-trained audio model, and fires a vibration motor on an ESP32 wristband over Wi-Fi. It is not meant to over-engineer the sport. It levels the field so a player can rely on instinct.
02 How it works
The mechanism.
- 01Continuous audio capture at 16 kHz
- 02YAMNet (TensorFlow Hub, trained on AudioSet) classifies whistle events
- 03Python backend sends an HTTP request over Wi-Fi to an ESP32 DevKit
- 04ESP32 drives a vibration motor on GPIO 21 for three seconds
- 05Tested in a controlled environment and on turf
Stack
- TensorFlow Hub YAMNet
- Python, NumPy, sounddevice
- ESP32, Arduino IDE
- Vibration motor, USB or battery wristband