EDSNet:信号数目未知场景下的时频混叠信号一体检测与分离方法OA

With the rapid development of technologies such as electromagnetic spectrum warfare,the electromagnetic environment is becoming increasingly complex,and the problem of various types of signal aliasing has become prominent.Conventional separation of aliased signals typically requires prior knowledge of the number of signals,making it difficult to adapt to unknown open electro-magnetic environments.To address the above issues,an integrated detection and separation single-channel deep learning model is proposed in this paper,namely the electromagnetic detection-separation network.Firstly,a high-dimensional feature processing ar-chitecture is designed to achieve representation of aliased signals.Secondly,by fusing multi-scale convolutions and lightweight at-tention mechanisms,a multi-kernel attention encoder is designed to enhance the feature representation capability for aliased sig-nals.Finally,a detection head and a mask generation module are designed.The detection head is utilized to estimate the number of signal components contained in aliased signals,and the mask generator adaptively selects the corresponding mask generation u-nits according to the number of signals to achieve reconstruction and separation of each component signal.Simulation experiments show that the proposed method can still effectively achieve signal separation when the number of aliased signals is unknown.