Compressive radar clutter subspace estimation using dictionary learning

Abstract

Abstract-Space-Time Adaptive Processing (STAP) based on matched filter processing in the presence of additive clutter (modeled as colored noise) requires knowledge of the clutter covariance matrix. In practice, this is estimated via the sample covariance matrix using samples from the neighboring range bins around the reference bin. By applying compressive sensing, the number of training samples needed to estimate the covariance matrix can be significantly reduced, provided that the basis mismatch problem, inherent to compressive sensing can be mitigated. This paper presents an adaptive approach to choosing the best sparsifying basis, using dictionary learning to estimate the radar clutter subspace. Numerical results show that the proposed algorithm achieves the desired reduction in training samples, and is more accurate than previous reduced-rank algorithm baseline.

Sparsity-Based STAP Design Based on Alternating Direction Method with Gain/Phase Errors

“…That is to say the clutter sparsity is much lower than the system DoFs and far lower than N d N s (since the clutter rank is much lower than N M and N M ≫ N d N s ). Similar conclusions are also obtained by L. Bai [77]. Moreover, according to [8], the clutter rank can be estimated by counting the number of resolution grids that are occupied by the significant clutter spectrum components.…”

Discrete Interference Suppression Method Based on Robust Sparse Bayesian Learning for STAP

“…denotes the set of spatial-temporal steering vectors covering the region of interest, andÑ decides the number of the spatial-temporal steering vectors for constructing the block matrix. Because clutter may not be exactly located exactly on a gird point of the dictionary, thus in order to further improve the estimation performance, the off-grid mismatch between the space-time overcomplete dictionary and the actual clutter distribution needs to be considered [17], [34], [35]. In the proposed method, the space-time snapshot of the range cell under test is rewritten asx…”

“…As such, the overcomplete dictionary can be calibrated iteratively by minimizing the cost function, thus the error between the dictionary and the real clutter distribution is reduced effectively. Compared with the mismatch calibration method in [35], the constraint conditions of mismatch calibration in the proposed method is less, the proposed method does not need to set the range of mismatch parameters, thus it is much more convenient to be applied.…”

Robust STAP algorithm based on knowledge‐aided SR for airborne radar

“…Some distinct elements in α correspond to outliers in this scenario. Especially, when outliers are dense in training snapshots, SR‐STAP algorithm has obvious error in the CCM estimation [24], which leads to target self‐nulling phenomenon. Hence, the prior knowledge about clutter distribution described as (2), which can distinguish clutter components and outliers effectively, is exploited to improve the robustness of SR.…”