Coherent adaptive radar detection in non-Gaussian clutter

Adaptive detection of fluctuating radar targets in unknown correlated Gaussian disturbances has received considerable attention in the past. Kelly's (1986) generalized likelihood ratio test (GLRT) is the preferred algorithm for detecting Swerling-I targets in Gaussian noise. However, the problem of adaptive detection in non-Gaussian environment is still under investigation. In this paper we pursue two aims: to investigate the performance of Kelly's GLRT in non-Gaussian clutter; (ii) to derive a detection algorithm with constant false alarm rate (CFAR) behavior with respect to the amplitude probability density function (apdf) parameters and to the correlation structure of the disturbance that outperforms Kelly's GLRT in non-Gaussian clutter. Performance analysis is presented using both simulated data and real sea clutter data.