Multi-network Based Automatic Modulation Recognition with Confidence Fusion

Automatic modulation recognition (AMR) is of great importance in various communications applications. Deep Learning (DL) based AMR has been becoming a popular choice as it inhibits the powerful classification capability of DL. However, the existing DL based AMR methods mostly exploit a single DL network and can hardly obtain a full understanding of the received signal. This paper proposes a multi-network based AMR algorithm with confidence fusion to improve the recognition performance. According to the proposed algorithm, four different deep neural networks (DNNs) are adopted to handle the received signal individually, and confidence fusion is used to combine the outputs of multiple networks to produce the final decisions. Simulation results show the superiority of the proposed algorithm in recognition accuracy.