Improved Neyman-Pearson Network for MIMO Radar Moving Target Detection

Abstract

In this paper, we consider the moving target detection problem for distributed multiple-input multiple-output (MIMO) radar in compound-Gaussian clutter environment. By introducing a binary discrete variable, the problem of target detection is transformed into the estimation problem of discrete variable, and on the basis of Neyman-Pearson network (NPnet), an Improved Neyman-Pearson network (INPnet) is constructed to solve this problem. The INPnet modifies the nonlinear activation function of the output layer and loss function of the original network, and inherits the advantage of the original network combining data-driven and model-driven. In the network training stage, a training method considering the decision threshold is proposed, which achieves controllable false alarm rate. The simulation results show that the proposed detection method INPnet outperforms the existing NPnet, and can guarantee a constant false alarm rate (CFAR).