Sovereign 8s.is mirror of standardgalactic/dc-bias-optimization-lifi-and-validation
  • Python 39.8%
  • MATLAB 36.6%
  • C++ 23.6%
Find a file
Repository files (latest commit first)
Filename Latest commit message Latest commit date
2026-04-15 19:08:36 +02:00
data generation Add files via upload 2026-04-15 13:52:59 +02:00
hardware validation Delete hardware validation/n 2026-04-15 13:27:00 +02:00
ML models Delete ML models/m 2026-04-15 13:26:33 +02:00
README.md Fix formatting issues in README instructions 2026-04-15 19:08:36 +02:00

DC Bias Optimization in DCO-OFDM Li-Fi Systems

#Authors

This work is conducted by:

  • Esraa Abdelhakim
  • Dina A. Ragab
  • Mohamed Abaza
  • Mohamed Hussien Moharam

This repository contains the implementation associated with a research manuscript currently under review.

Project Overview

The goal of this work is to optimize DC bias in DCO-OFDM Li-Fi systems using hybrid machine learning techniques. The study includes:

  • Simulation-based dataset generation (MATLAB)
  • Appply different machine learning models (Python)
  • Real-time validation using Li-Fi hardware (Arduino + Photodiode)

System Flow

  1. MATLAB Simulation

    • Generate DCO-OFDM signals
    • Create dataset for ML models
  2. Machine Learning Models (Python)

    • Train multiple models
    • Compare performance
    • Select best model
  3. Hardware Implementation (Arduino)

    • Transmit data using LED (Li-Fi)
    • Receive data using photodiode
    • Extract signal features (Mean, Min, Max, Std, BER, Bias)
  4. Validation

    • Apply ML models on real hardware data
    • Verify consistency with simulation results

Repository Structure


How to Use

1. Generate Dataset

  • Run the MATLAB script:

2. Train ML Models

  • Run Models:

3. Hardware Experiment

  • Upload transmitter code to Arduino (LED side)
  • Upload receiver code to Arduino (photodiode side)
  • Collect output via Serial Monitor
  • Save data as CSV for ML processing

4. Test Hardware Dataset on ML Models

  • Load the hardware-generated dataset (CSV)
  • Apply the same trained ML models