SNR-Boosted Automatic Modulation Classification
Clayton A. Harper, Avi Sinha, Mitchell Aaron Thornton, Eric C. Larson · 2021 55th Asilomar Conference on Signals, Systems, and Computers · 2021
Automatic modulation classification is a desired feature in many modern software-defined radios; however, classification performance degrades with decreasing signal to noise ratios. We propose employing a deep convolutional signal to noise ratio estimation model to exploit relationships within signals of similar signal to noise ratio ranges through signal to noise ratio specific modulation classifiers. We utilize a two-stage process where the signal to noise ratio is first estimated and then demultiplexed into a modulation classifier that has been tuned on signals with similar signal to noise ratios. Using the proposed method, we build upon the current state-of-the-art and increase classification performance at decreasing signal to noise ratios.