Semi-Supervised Radio Signal Identification

Radio recognition in complex 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, but labeled data is often scarce making supervised learning strategies difficult and time consuming to curate. We demonstrate that semi-supervised learning techniques can be used to scale learning beyond supervised datasets, allowing for both discerning and recalling radio signals of interest by using sparse signal representations based on both unsupervised and supervised methods for nonlinear feature learning.