How swandip7/5gnr-snr-estimation-cef works
This repository provides synthetic 5G NR SNR-estimation datasets, MATLAB generation scripts, and a Python model-training and evaluation pipeline. The principal workflow generates channel samples and receiver-side measurements, stores features, labels, and estimator metadata in CSVs, then trains and evaluates fusion and baseline models. The README documents generator stages and capabilities; the sampled Python excerpt shows dataset loading and training components, but not the MATLAB internals or all training steps.
Main entry points: train_cef.py Training Pipeline · generate_dataset.m RF Impairments · Enhanced_5G_Dataset_OOD_5K_v311.csv OOD Dataset · Enhanced_5G_Dataset_100K_v311.csv Main Dataset
Components 16
Dataset Generation
- System Setup
matlab/generate_dataset.m - Scenario Selection
matlab/generate_dataset.m - CDL Channel
matlab/generate_dataset.m - MIMO Propagation
matlab/generate_dataset.m - RF Impairments
matlab/generate_dataset.m - Receiver Processing
matlab/generate_dataset.m - Feature Extraction
matlab/generate_dataset.m - Classical Estimators
matlab/generate_dataset.m
Datasets
- Main Dataset
data/Enhanced_5G_Dataset_100K_v311.csv - Main Metadata
data/Enhanced_5G_Dataset_100K_v3_metadata11.csv - OOD Dataset
data/Enhanced_5G_Dataset_OOD_5K_v311.csv - OOD Metadata
data/Enhanced_5G_Dataset_OOD_5K_v311_metadata.csv
Training and Evaluation
- Training Pipeline
python/train_cef.py - Model Architectures
python/train_cef.py - Evaluation and Ablations
python/train_cef.py
Other components
- Researcher
Connections 18
- Researcher → System Setup: generates data
README.md - System Setup → Scenario Selection: sets parameters
README.md - Scenario Selection → CDL Channel: selects profile
README.md - CDL Channel → MIMO Propagation: propagates signal
README.md - MIMO Propagation → RF Impairments: adds impairments
README.md - RF Impairments → Receiver Processing: feeds receiver
README.md - Receiver Processing → Feature Extraction: derives features
README.md - Receiver Processing → Classical Estimators: computes estimates
README.md - Feature Extraction → Main Dataset: stores samples
README.md - Classical Estimators → Main Metadata: stores estimates
README.md - Researcher → Training Pipeline: trains models
README.md - Main Dataset → Training Pipeline: loads samples
python/train_cef.py - OOD Dataset → Training Pipeline: loads OOD samples
python/train_cef.py - Training Pipeline → Model Architectures: fits models
python/train_cef.py - Training Pipeline → Evaluation and Ablations: evaluates results
python/train_cef.py - Researcher → OOD Dataset: generates OOD set
README.md - RF Impairments → OOD Dataset: shifts severity
README.md - Classical Estimators → OOD Metadata: stores estimates
README.md