Automatic Modulation and Recognition of Robot Communication Signal Based on Deep Learning Neural Network

In order to solve the problem that traditional method of manually extracting expert features for communication signal recognition has large limitations and low accuracy under signal-to-noise ratio, this paper proposes an automatic modulation robot based on deep learning neural network. method, received is preprocessed obtain complex baseband including in-phase component quadrature component. The used as data set input convolution network model. model structure super parameters such kernel, step size, characteristic graph, activation function are adjusted through multiple training, trained extract recognize signal. It realizes identification classification seven types digital signals: 2FSK, 4FSK, BPSK, 8PSK, QPSK, QAM16, QAM64. experimental results show average signals reached 94.61% when ratio 0 dB. Conclusion. algorithm proved be effective high condition ratio.