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.
Highlights
R ADAR emitter classification (REC) based on the passively received radar pulse streams is of great importance in both military and civil systems, especially in electronic support measurement (ESM) systems [1]
Experiments on real-world data streams collected in multi-emitter scenario prove the validity and superiority of dynamic improved exemplar-based subspace clustering (DI-exemplar-based subspace clustering (ESC)) compared with state-of-the-art data stream clustering (DSC) algorithms including scalable sparse subspace clustering (SSSC) [18], scalable low-rank representation (SLRR) [18], scalable least squares regression (SLSR) [18], clustering of evolving data streams into arbitrary shapes (CEDAS) [19] and stream affinity propagation (STRAP) [20]
We propose two clustering algorithms, called improved ESC (I-ESC) and DI-ESC, which can deal with the subspace-imbalanced data sets as well as subspace-imbalanced and evolving data streams, respectively
Summary
R ADAR emitter classification (REC) based on the passively received radar pulse streams is of great importance in both military and civil systems, especially in electronic support measurement (ESM) systems [1]. The researchers focus on solving the REC problem through supervised learning approaches They assume that the pulse samples of all radar emitters can be pre-obtained and used to train delicate classifiers for REC purposes. Despite the pulses are high-dimensional, it has been found that the radar pulses from the same emitters lie on or near low-dimensional subspaces [15], [16] From this sense, the REC task can be deemed to perform an online and incremental subspace clustering on the pulse stream whose goal is to group the pulses from the same emitters together and separate the pulses from the different emitters. A data stream subspace clustering algorithm, called dynamic improved ESC (DI-ESC) algorithm, is proposed to achieve online REC on pulse streams, where the pulses from the same radar emitters are assumed being located in the same subspace. Experiments on real-world data streams collected in multi-emitter scenario prove the validity and superiority of DI-ESC compared with state-of-the-art DSC algorithms including scalable sparse subspace clustering (SSSC) [18], scalable low-rank representation (SLRR) [18], scalable least squares regression (SLSR) [18], clustering of evolving data streams into arbitrary shapes (CEDAS) [19] and stream affinity propagation (STRAP) [20]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
