Improving aerial target detection for 3D radar based on a two-stage CFAR method with adaptive clutter distribution estimation

This study deals with the problem of enhancing aerial target detection for 3D radar. A novel approach which incorporates both signal and data processing is introduced. In order to increase the target’s SNR (signal-to-noise ratio), two consecutive transmit beams are used; for each, three beams are received simultaneously. All received beams are then processed. A two-stage constant false alarm rate (CFAR) algorithm is proposed for improving target detection. At the first-stage CFAR, the global CA-CFAR is applied to identify all possible target candidates (plots). Then, unsupervised machine learning is used to separate interference regions. For each interference region, the truncated probability density function of interference is estimated, and then a local CFAR (second-stage CFAR) is applied to reduce false plots while retaining target plots. The proposed approach is an extension of that given in recent publications. Tests on a 3D surveillance radar show the effectiveness of the proposed approach on aerial target detection in comparison with previous methods.