The traditional pulse repetition interval (PRI) transform method and modified PRI transform method have a low ability to recognise the complex PRI modulation type adopted in spaceborne synthetic aperture radar systems. To solve this problem, a short-time modified PRI transform (STMPT) method based on the traditional PRI transform method was proposed. The STMPT method converts the one-dimensional representation of PRI estimation results into a time-varying characteristic sensitive (TVCS) matrix of the PRI. Combined with the mathematical models of the four PRI modulation types including PRI fixed, staggered PRI, fast PRI change, and a more elaborated PRI sequence concatenating multiple fast PRI change sequences, the nonzero segment characteristics were obtained by incoherent accumulation and rearrangement of the TVCS matrix. Parameter estimation of the four PRI modulation types was also realised based on the geometrical and statistical characteristics. Finally, the effectiveness and performance of the algorithm were verified by simulation analysis.

No abstract

No abstract

Most existing methods for sorting synthetic aperture radar (SAR) emitter signals rely on either unsupervised clustering or supervised classification methods. However, unsupervised clustering can consume a significant amount of computational and storage space and is sensitive to the setting of hyperparameters, while supervised classification requires a considerable number of labeled samples. To address these limitations, we propose a self-supervised clustering-based method for sorting SAR radiation source signals. The method uses a constructed affinity propagation-convolutional neural network (AP-CNN) to perform self-supervised clustering of a large number of unlabeled signal time-frequency images into multiple clusters in the first stage. Subsequently, it uses a self-organizing map (SOM) network combined with inter-pulse parameters for further sorting in the second stage. The simulation results demonstrate that the proposed method outperforms other depth models and conventional methods in the environment where Gaussian white noise affects the signal. The experiments conducted using measured data also show the superiority of the proposed method in this paper.