Adaptive Radar Detection of Dim Moving Targets in Presence of Range Migration

Abstract

This paper addresses adaptive radar detection of dim moving targets. To circumvent range migration, the detection problem is formulated as a multiple hypothesis test and solved applying model order selection rules which allow to estimate the "position" of the target within the CPI and eventually detect it. The performance analysis proves the effectiveness of the proposed approach also in comparison to existing alternatives.Index Terms-Adaptive radar detection, generalized likelihood ratio test, range migration, model order selection.

Adaptive Detection of Dim Maneuvering Targets in Adjacent Range Cells

“…Moreover, denote by N p pulses belonging to the transmitted pulse train. Then, for a point-like target, the signal received by the mth antenna element can be written as [14] x m (t) = α…”

“…In [13], an alternative method relying on adjacent correlation function and Lv's transform is devised to detect the maneuvering targets with radial jerk motion. More recently, in [14], innovative one-step and twostep detection architectures are conceived for dim maneuvering targets with and without estimating the slow-time index of the target signal in the CUT and based upon the generalized information criterion [15]. Remarkably, such architectures can overcome conventional detectors as the generalized adaptive matched filter (GAMF) [16] at the price of an increased computational complexity.…”

“…In this letter, we focus on the detection of dim maneuvering targets in the presence of RCM and further improve the results of [14] by devising innovative robust (with respect to the amount of training samples) architectures. To this end, we do not consider any possible phase/amplitude relationships between consecutive pulses and exploit, at the design stage, the Bayesian information criterion (BIC) rule [15], which is an asymptotic approximation of the optimal maximum a posteriori rule, to identify the pulse echoes containing target components over two consecutive range cells.…”

“…Then, we conceive two-step architectures (TSA) and one-step architectures (OSA) relying on GLRT-based design criteria, where GLRT stands for generalized likelihood ratio test. The contributions of the present letter can be summarized as follows: 1) unlike [14], all the samples from two consecutive range cells occupied by the target are processed to increase the detection performance; 2) the samples free of signal components are exploited for the estimation of the interference covariance arXiv:2103.04367v1 [eess.SP] 7 Mar 2021 matrix (ICM) lending new architectures a robustness to the training set size; 3) the proposed architectures are designed to avoid a continuous computation of inverse matrices saving computational resources.…”

Radar Clutter Classification Using Expectation-Maximization Method

“…In the context of the design of adaptive radar detection architectures, it is usually assumed that a set of secondary data, which are target-free and share the same statistical properties of the interference as the primary data (homogeneous assumption) is available [1][2][3][4][5]. The secondary data are typically collected in the proximity of the range cell under test and are used to achieve adaptivity as well as to ensure the Constant False Alarm Rate (CFAR) property.…”