Compound Jamming Signal Recognition Based on Neural Networks
Fu Ruo-Ran · 2016
An algorithm of recognizing radar compound jamming signals including additive, multiplicative and convolution signals of typical blanket jamming and deception jamming based on neural networks is proposed in this article. Firstly, all signals of echo, jamming and noise received in one pulse repetition interval are acquired as signal sources. Then the features of the signal sources are extracted in time domain, frequency domain and fractal dimensions. Finally, classifier based on neural networks is established, by which compound signals are recognized. Results of the experiment indicate that the algorithm has the ability to recognize not only compound modes but also signal types, which enhances the accuracy of recognition.