A Study Of Adaptive Constant False Alarm Rate Detection

In the radar system, constant false alarm rate is one of the most important issues that each radar designer should pay attention to. And this technique is the most important means to control the false alarm rate in the radar target detection systems.Based on the deep understanding of the target detection theory and the constant false alarm rate algorithm, we research on the adaptive CFAR detection algorithm in the Gaussian clutter background which includes homogeneous case, clutter edge heterogeneous situations and multiple targets situations.In this paper, the main CFAR algorithms consist of the mean level type CFAR, order statistics type CFAR, and clutter map CFAR. For the mean level CFAR algorithm, CACFAR, GOCFAR and SOCFAR are considered in detail. The fundamental principles and the implementation procedures of these algorithms are introduced simply. Then the characteristics of each CFAR detectors are analyzed in the homogeneous clutter background,the clutter edge heterogeneous and the multiple targets cases. Next, the detection results in different clutter backgrounds are given through the computer simulations, and the CFAR losses are also discussed.When the clutter varies seriously in the spatial domain and is stationary in the temporal domain, the clutter map algorithm will be preferred to estimate the clutter characteristics. The detection performance for the low and high velocity targets of the point clutter map CFAR detector and the surface clutter map CFAR detector are analyzed, besides the merits and drawbacks are also discussed. Both of the simulation and the experimental data are processed in this paper. The experimental results illustrate the features of each detector.Finally, an adaptive CFAR detector selection method is proposed based on the mean level CFAR detector and the clutter map. In the proposed method, the clutter map of the area under test is constructed firstly. Depending on the clutter map, the distribution of the clutter is judged to distinguish the homogeneous clutter area, weak clutter area and strong clutter area of the clutter edge. Then, the most appreciate CFAR detector isselected according to the distribution of the clutter. The results show that the proposed CFAR detector selection method can efficiently improve the detection performance in the complicated clutter background.