Performance Analysis of Non Coherent CFAR Detection Based on Goodness-of-Fit Tests in Different Clutter Environments

Non-coherent detection using multiple pulses is very useful in radar detection, especially in the case that the phase information of return pulses is unavailable. This paper addresses the performance of non-coherent CFAR detection based on goodness-of-fit (GoF) tests compared with non-coherent CFAR integrator (NCCI) in different clutter environments by simulation experiments, when the assumed and actual clutter distributions are same and different respectively. The results show that the GoF CFAR detector outperforms the NCCI in very spiky Weibull, lognormal as well as K clutter with enough integrated pulses, even when the assumed and actual clutter distributions are different. The CFAR loss of the GoF CFAR detector due to the estimation of unknown distribution parameters has also been shown in this paper.