Distributed Automatic Modulation Classification Based on Cyclic Feature via Compressive Sensing

Automatic modulation classification (AMC) is an important component in cognitive radio and many efforts have been made to improve the AMC's successful classification rate, especially when the environment is noisy. The cyclic feature has excellent resiliency to noise, so it has been frequently adopted as the feature for AMC. In this paper, in order to enhance the reliability of the system, we propose a distributed AMC scheme based on compressive sensing by taking advantage of the sparse property of cyclic feature map. A novel method based on compressive sensing principle for capturing the prominent peaks of the feature map is introduced, which is able to achieve good modulation recognition performance at sub-Nyquist rates. And a novel neural network fusion strategy is proposed for better cooperation. It is shown that the proposed distributed approach improves the classification rate compared with the single radio and can reduce the required samples compared with traditional sampling scheme.