Design of Radar Target Detector based on CFAR Feature Plane

The processing framework resorting to the constant false alarm rate (CFAR) feature plane interprets CFAR detector design as a learning machine composed of input layer, hidden layer (invariant feature extraction), design layer and output layer, which provides a novel way to the design of CFAR detector. In this paper, we investigate the design criterion to the design layer under this framework. On the CFAR feature plane composed of two maximum invariant statistics, two new detectors with simple controllable parameters are designed through feature fusion by weighted arithmetic mean and weighted geometric mean. Finally, simulation experiments verify the effectiveness of the two methods.