Signal denoising recognition of shipborne phased array radar based on high-order time-frequency spectrum features
Aiming at the problem that the noise signal spectrum energy of shipborne phased array radar signals exceeds the target signal in complex electromagnetic environments,making it difficult to recognize radar signals,a denoising recog-nition method for shipborne phased array radar signals based on high-order time-frequency spectrum features is proposed.By using the CWD time-frequency transformation method,the original radar signal is mapped to the time-frequency domain.After extracting the two-dimensional time-frequency distribution information of the signal,the CWD time-frequency distri-bution information is power processed to calculate the power information of each element in the time-frequency information and extract the high-order time-frequency spectrum characteristics of the signal;By using the generalized S-transform meth-od,the Gaussian window function and time-frequency resolution used for filtering high-order time-frequency spectrum fea-tures of the signal are adjusted to achieve time-frequency denoising of radar signals.Input the denoised signal into a recur-rent neural network,capture the temporal dependencies in the denoised signal,learn the distinguishing features between dif-ferent signal types,and complete radar signal denoising recognition.After testing,this method has shown accurate recogni-tion results for signals under strong electromagnetic interference and can effectively weaken the impact of complex electro-magnetic environment interference.
high order time-frequency spectrum featuresshipborne phased arrayradar signalnoise reduction re-cognitionrecurrent neural networkgeneralized S-transform
NETL
NSTL
万方数据
