Radar Emitter Signal Recognition based on Contour Lines of Ambiguity Function and Stacked Denoising Auto-encoders
Yunwei Pu, Jiang Hong Guo, Can Liu, Haixiao Wu · 2021
Aiming at many problems of complex radar emitter signal recognition methods, like poor anti-noise performance and low recognition rate. In this paper, a new recognition method based on contour lines of ambiguity function and stacked denoising auto-encoders is proposed. First, an ambiguity function is processed by the Gaussian smoothing and calculate contour lines by linear interpolation. Then, principal component analysis is used to reduce its feature dimension, retain the main ambiguity energy information. Finally, deep learning stacked denoising auto-encoders are built to learn and extract the deep and more ubiquitous features of contour lines, and classify them through the Softmax classifier. The simulated experiments show that the overall average recognition rate of the six typical radar signals is maintained above 99.83% when the signal-noise ratio is 0dB. Even in the -6dB environment, it also can reach 83.67%, which proves this method has good performance and feasibility under the extremely low signal-noise ratio conditions.