CFAR detector based on clutter partition in heterogeneous background

The conventional CFAR detectors based on reference window suffer a considerable performance degradation in heterogeneous background. In this paper, a new knowledge based constant false alarm rate (KB-CFAR) detector is proposed to reduce the impact of heterogeneous environment. The proposed KB-CFAR detector exploits clutter partition as auxiliary knowledge to select the reference cells. The clutter partition is obtained by extracting useful knowledge from the massive environment information and it is updated by real-time radar returns. The performance of the new KB-CFAR detector is analyzed by real radar data which is collected by a linear frequency-modulated continuous wave (LFMCW) radar. The results show that the new detector achieves a satisfactory performance and works well under general conditions.