DNN-based Intelligent Target Detection for MIMO Radar

This paper investigates intelligent target detection for Multiple-Input Multiple-Output (MIMO) radar in complex clutter and noise environments. In the modeling phase, the detection problem is formulated as a binary hypothesis testing scenario. To handle the detection problem, adaptive detectors based on Deep Neural Network (DNN) are developed resorting to the Generalized Likelihood Ratio Test (GLRT) and Adaptive Matched Filter (AMF) criteria by constructing a network architecture comprising input layers, multiple hidden layers, dropout layers, and output layers. Specifically, a supervised learning method based on DNN is developed, where the DNN is represented as a parametric function, mapping inputs to outputs through consecutive hidden layers. In the analysis phase, the proposed DNN-based detectors are compared with traditional GLRT and AMF-based counterparts. Numerical findings confirm the efficiency of the devised DNN-based strategies.