Classification of low probability of interception communication signal modulations based on time-frequency analysis and artificial neural network

Classification of modulation types faces a problem of low SNR in conditions where Low Probability of Interception signals are used. A novel feature vector extraction algorithm fit for LPI communication signals is presented, in which feature vector is generated by autonomously cropping the modulation energy from Time-Frequency images. Multi-Layered Perceptron is adopted as classification decision parts. Probabilities of correct classification are obtained via computer simulation. The results show that the classification scheme proposed in this paper has promising performance in low SNR conditions.