Radar modulation recognition based on MLP neural network
Fang Li, Yu Wang, Liangliang Zhao, Zixian Yang · 2019
In this paper, a radar modulation recognition method based on improved multi-layer perception (MLP) neural network is proposed. Some improvements are made in following aspects: initialization of the network model weights, normalizetion of the input parameters and setting of the learning rate. Simulation experiments show that this method has high recognition and generalization capability with the relatively simple network model. When the signal-to-noise ratio is as low as 0 dB, the correct recognition rate is over 90%, and meanwhile it has good recognition performance for complex intra-pulse modulation signals.