Abstract :
The performance of a radar system is determined in part by the detector\´s ability to isolate returns occurring from desired targets from those that are originating from clutter and noise. When the requirement is to detect targets against a homogeneous background, Cell Averaging (CA) CFAR provides the optimum performance. However, there is a significant degradation when the assumption of homogeneous environment is not met. In these non-uniform clutter environments the Order Statistic (OS) CFAR is preferred but at the cost of a moderate degradation in detection probability. In a coherent radar system the dynamics of the target also plays a critical part in determining if the target is detected against either a clutter or noise-like background. Optimal detection of targets with different dynamics requires different CFAR schemes and parameters. Optimal performance therefore requires utilizing variant CFAR\´s in a hierarchical manner as well as making the parameters of the CFARs adaptive. This is the logic diagram of a Switched Mode CFAR that uses a clutter masking approach to switch between CA CFAR and its variants and OS CFAR. These algorithms also include adaptive footprints to better estimate the test cells background distributions in mixed clutter environments and to suppress the false alarm breakthrough due to clutters. This results in an Adaptive Switching (AS) CFAR. Performance improvements are demonstrated with data from a HF Surface Ware Radar (HFSWR). The analysis includes the processing of twenty four hours of data with both AS CFAR and a conventional detector with each output feeding the common tracker. The trackers outputs are analyzed and compared. [From: "An Adaptive Hierarchal CFAR for Optimal Target Detection in Mixed Cluner Environments," Radar 2011, 3179.]
