M.Tech AI Research Project

GNSS Error Prediction Dashboard

Hybrid LSTM-Transformer for time-varying satellite clock and ephemeris error analysis.

Problem

Predict GNSS broadcast-model mismatch errors across multiple validity windows.

Models

LSTM baseline, Transformer, and Hybrid LSTM-Transformer architecture.

Output

RMSE, MAE, R², prediction plots, and error-distribution analysis.

Sample Results

These plots are generated by the training pipeline and stored in the repository.

Run the Full Web App

GitHub Pages hosts static content only. For the interactive Flask dashboard, run locally with Python or deploy backend to Render/Railway.

python src/web_app/app.py