Using signal processing based features to train and validate machine-learning algorithms to improve spectrum sensing and related problems in cognitive radios.
-
Updated
Feb 10, 2023 - Jupyter Notebook
Using signal processing based features to train and validate machine-learning algorithms to improve spectrum sensing and related problems in cognitive radios.
Spectrum sensing in cognitive radios leveraging machine learning models
In this repository, we deal with developing an energy detector and a detector based on cyclostationarity for an OFDM based cognitive radio system and implementing and evaluating the performance of these detectors.
Code and resources for the paper: "Cognitive Radio Spectrum Sensing on the Edge: A Quantization-Aware Deep Learning Approach"
Exploring Rayleigh fading channels for NOMA users, our project uses Monte Carlo simulations to analyze signal detection across various SNRs.
TinyML-aware spectrogram segmentation using pruning and knowledge distillation for RF spectrum sensing.
Energy and polarization based interference mitigation
Cognitive passive RF spectrum sensing & Specific Emitter Identification (SEI) on HackRF/SDR — multi-band detection (BT/BLE/Wi-Fi/ZigBee/LoRa/LTE/radar), explainable & reproducible.
A curated, task-oriented catalogue of datasets and supporting resources for machine learning research in wireless communications.
Angular Domain-Based Cyclostationary Feature Detector for Spectrum Sensing
Master's thesis (FAU): RSS-based transmitter localization (MSE & LIvE methods) and radio environment map construction via distributed spectrum sensing for 5G, with a 5G-NR-style OFDM waveform extension and spatially correlated shadowing modeling in MATLAB.
Cognitive radio simulation in C++: dynamic common control channel with Markov PU prediction, underlay/interweave access, and backoff-based displacement recovery (with a verified spectrum-sensing baseline)
Reproducible artifact for the ICE2CT-2026 paper Dual-Stream Phase-Aware Inception-Time CNN for adversarially robust V2X spectrum sensing: CPU-only dual-stream CNN (86k params), FGSM/PGD/APGD/FAB/Square evaluation, corrected metrics (robust accuracy + conditional ASR), one-command reproduction, and a full honesty audit of legacy results.
To associate your repository with the spectrum-sensing topic, visit your repo's landing page and select "manage topics."