Semi-Supervised Radio Signal Identification

TL;DR

This paper introduces semi-supervised learning methods for radio signal identification, enabling recognition of new signals with limited labeled data by combining unsupervised and supervised feature learning and clustering.

Contribution

It presents a novel semi-supervised approach that leverages sparse signal representations to improve radio emitter recognition in dense environments.

Findings

Semi-supervised techniques outperform purely supervised methods.

Sparse signal representations facilitate better clustering of radio signals.

The approach enables recognition of new signals with minimal labeled data.

Abstract

Radio emitter recognition in dense multi-user environments is an important tool for optimizing spectrum utilization, identifying and minimizing interference, and enforcing spectrum policy. Radio data is readily available and easy to obtain from an antenna, but labeled and curated data is often scarce making supervised learning strategies difficult and time consuming in practice. We demonstrate that semi-supervised learning techniques can be used to scale learning beyond supervised datasets, allowing for discerning and recalling new radio signals by using sparse signal representations based on both unsupervised and supervised methods for nonlinear feature learning and clustering methods.