Online Non-Cooperative Radar Emitter Classification From Evolving and Imbalanced Pulse Streams

Abstract

Recent research treats radar emitter classification (REC) problems as typical closed-set classification problems, i.e., assuming all radar emitters are cooperative and their pulses can be pre-obtained for training the classifiers. However, such overly ideal assumptions have made it difficult to fit real-world REC problems into such restricted models. In this paper, to achieve online REC in a more realistic way, we convert the online REC problem into dynamically performing subspace clustering on pulse streams. Meanwhile, the pulse streams have evolving and imbalanced properties which are mainly caused by the existence of the non-cooperative emitters. Specifically, a novel data stream clustering (DSC) algorithm, called dynamic improved exemplar-based subspace clustering (DI-ESC), is proposed, which consists of two phases, i.e., initialization and online clustering. First, to achieve subspace clustering on subspace-imbalanced data, a static clustering approach called the improved ESC algorithm (I-ESC) is proposed. Second, based on the subspace clustering results obtained, DI-ESC can process the pulse stream in real-time and can further detect the emitter evolution by the proposed evolution detection strategy. The typically dynamic behavior of emitters such as appearing, disappearing and recurring can be detected and adapted by the DI-ESC. Extinct experiments on real-world emitter data show the sensitivity, effectiveness, and superiority of the proposed I-ESC and DI-ESC algorithms.

Dynamic Sparse Subspace Clustering for Evolving High-Dimensional Data Streams

“…where r j (•) is the residual if x t is assigned into the jth subspace. The optimal j * will be obtained by (18) which corresponds to the j * th subpart in Z t . If j * ≤ k t , the corresponding active subspace S j * would be updated, otherwise the inactive subspace D j * −k t would be updated.…”

“…Then, the point will be accepted as a normal point or rejected as an outlier according to (17). For normal points, they will be updated in the EDSSC summary or the remove reservoir by (18). While the outliers will be put into the outlier reservoir O t (Section IV-B).…”

“…Motivations: It has been realized that many real-world high-dimensional data, such as motion [11], face [12], and texture [13], actually lie in a union of low-dimensional subspaces [1], [2], [14]- [18]. Subspace clustering (SC) refers to the problem of simultaneously clustering the data into multiple subspaces and finding a low-dimensional subspace to fit each group of points [19]- [22].…”

Radar Signal Deinterleaving in Electronic Warfare Systems: A Combined Approach

“…PDWs include parameters such as time of arrival (TOA), frequency (RF), pulse width (PW), and angle of arrival (AOA). If the signals are sufficiently separated from each other in the RF-PW-AOA parameters domain, a simple clustering algorithm can be successful without the TOA analysis [3][4][5][6][7][8][9][10]. However, radars can have intersecting parameter ranges, and some of the parameters might be missing in the PDWs.…”

Improvement of Starting Point Selection of Data Field Clustering Algorithm

“…Radar emitter signal sorting, as a key step of electronic reconnaissance, refers to classifying the mixed signals received in the battlefield environment from different emitters to different clusters according to radar pulse parameter, so as to facilitate the subsequent pulse accumulation and realize deeper target identification and threat determination [1][2]. However, in the modern battlefield, with the explosive growth of radar equipment, the radar signal density has reached millions per second in the same space, and the continuous application of anti-reconnaissance technology makes the radar signal form more complex and changeable, which brings great challenges to the radar signal sorting technology [3][4][5].…”