Persymmetric GLRT-Based Detectors with Training Data for FDA-MIMO Radar

In the context of frequency diversity array multiple-input multiple-output (FDA-MIMO) radar employing symmetrically spaced linear transmit and receive arrays, the noise covariance matrix exhibits a persymmetric characteristic. Exploiting this prior knowledge of the covariance matrix structure, this paper tackles the challenge of detecting a moving target against a Gaussian background using FDA-MIMO radar. Grounded on the one-step and two-step generalized likelihood ratio test (GLRT) criteria - OGLRT and TGLRT, respectively - two adaptive detectors are developed utilizing training data. Additionally, analytical expressions for the detection probability (PD) and false alarm probability (PFA) of these detectors are derived, revealing their constant false alarm rate (CFAR) property relative to the covariance matrix. Numerical simulations underscore the advantages of these detectors, demonstrating significant improvements in detection performance and reducing the amount of required training data. Moreover, an effective method is provided to enhance the alignment between theoretical and simulated PD outcomes for the OGLRT-based detector under conditions of limited sample availability.