Skip to content

phlangone/esp-somgoritmo

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Wearable Device for Sound Hazard Detection

Assistive Technology for Deaf and Hard-of-Hearing Individuals


Overview

image

This project presents a wearable embedded system designed to assist deaf and hard-of-hearing individuals in detecting critical sound-based events in their environment.

The solution combines embedded systems, real-time operating systems (RTOS), and cloud-based machine learning inference to provide reliable hazard detection and user feedback.

The system focuses especially on industrial and urban safety scenarios, where sound awareness is crucial.


Social Context

  • ~865,000 Brazilians with hearing impairment in working age
  • Only ~8% are formally employed
  • Industrial environments present additional safety challenges

This project explores how AI + Embedded Systems can bridge this gap through assistive technology.


System Architecture

Server-Based Inference Architecture

image
  • Lightweight embedded device
  • Audio processing + transmission
  • Cloud-based neural network inference
  • Fast feedback to the user

Embedded Side (Device)

  • Captures environmental audio
  • Performs signal conditioning
  • Sends data to the server via WiFi
  • Receives classification results
  • Notifies the user

Cloud Side (Server)

  • Receives audio samples
  • Applies signal processing pipeline
  • Extracts features (MFCC)
  • Performs classification using CNN
  • Returns hazard label

Hardware Components

image
  • Microcontroller: ESP32-WROOM
  • Microphone: INMP441 (I2S MEMS)
  • Display: SSD1306 OLED (I2C)
  • Actuator: Vibration Motor
  • Connectivity: WiFi

Firmware Architecture

image

The firmware is structured using FreeRTOS, enabling modular and concurrent execution:

Main Tasks

  • Audio Capture Task
  • Signal Processing Task
  • Communication Task (HTTP/MQTT)

Signal Processing Pipeline

image

1. FIR Filtering

  • Order: 100
  • Window: Hamming
  • Noise reduction

2. Feature Extraction

  • MFCC (Mel Frequency Cepstral Coefficients)
  • Efficient representation for classification

Machine Learning

image
  • Model: CNN
  • Dataset: UrbanSound8K

Training Setup

  • 80 epochs
  • Categorical Cross-Entropy
  • Adam optimizer

User Feedback

image
  • Vibration motor → tactile alert
  • OLED display → hazard identification

How It Works

  1. Audio capture
  2. Transmission to server
  3. Signal processing + MFCC
  4. CNN classification
  5. Response to device
  6. User alert

Limitations & Trade-offs

  • Network dependency
  • Latency
  • Security considerations
  • Trade-off vs TinyML

Future Improvements

  • Edge inference (TinyML)
  • Hybrid architecture
  • Latency optimization
  • Industrial validation

Contact


License

Educational and research purposes

About

Wearable assistive device for hearing-impaired users using neural network-based audio recognition to detect environmental hazards.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors