Voice Recognition Device for Immediate Assistance
Software as a Service
NoCategory
Thesis DesignClients
Techstack
Purpose
To build a device that recognizes the voice signals of elders and infants and immediately alerts caregivers through a companion mobile app — combining Raspberry Pi audio capture, cloud speech recognition, and push notifications into one assistance system.
Description
A microphone-equipped Raspberry Pi continuously captures audio with sounddevice and performs Automatic Speech Recognition through the Google Cloud Speech-to-Text API. Recognized assistance calls update Firebase Realtime Database and trigger Firebase Cloud Messaging push notifications to a React Native mobile app, so caregivers and parents are alerted the moment help is needed. The device runs as a systemd service with GPIO status control, and the study evaluated the recognition accuracy and overall effectiveness of the system.
