Digital Operating Mode Classification of Real-World Amateur Radio Transmissions

2025-01-13

AI Summary

This work addresses the challenge of digital signal mode recognition in amateur radio. We propose the first fine-grained classification system trained on real-world UHF-band measurements, covering 17 digital modes and 98 parametrized signal variants. To mitigate non-ideal channel effects, we design an online spectrogram enhancement pipeline that jointly models multipath fading, Doppler shift, and phase noise—implemented via a dual-SDR receiver architecture. We further demonstrate, for the first time, the generalization capability of a lightweight EfficientNet-B0 model under realistic wireless channel conditions; remarkably, the model trained solely on random-payload data achieves robust performance on real Wikipedia-payload test signals. Evaluated on measured data, our system attains 93.80% accuracy at the mode level and 85.47% at the parameter level. We also systematically quantify the impact of signal duration, FFT resolution, and SNR on classification performance.

Technology Category

• Machine Learning: Multimodal Learning
• Planning, Routing, and Scheduling: Activity and Plan Recognition
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Responsible Web: Measurement, analysis, and circumvention of Web censorship

Abstract

This study presents an ML approach for classifying digital radio operating modes evaluated on real-world transmissions. We generated 98 different parameterized radio signals from 17 digital operating modes, transmitted each of them on the 70 cm (UHF) amateur radio band, and recorded our transmissions with two different architectures of SDR receivers. Three lightweight ML models were trained exclusively on spectrograms of limited non-transmitted signals with random characters as payloads. This training involved an online data augmentation pipeline to simulate various radio channel impairments. Our best model, EfficientNetB0, achieved an accuracy of 93.80% across the 17 operating modes and 85.47% across all 98 parameterized radio signals, evaluated on our real-world transmissions with Wikipedia articles as payloads. Furthermore, we analyzed the impact of varying signal durations&the number of FFT bins on classification, assessed the effectiveness of our simulated channel impairments, and tested our models across multiple simulated SNRs.

Problem

• Digital Radio Signal
• Classification
• Amateur Radio

Innovation

• Machine Learning Classification
• Digital Radio Signals
• EfficientNetB0 Model