Adaptive detection of a Gaussian signal in Gaussian noise

Adaptive detection of a Swerling I-II type target in Gaussian noise with unknown covariance matrix is addressed in this paper. The most celebrated approach to this problem is Kelly's generalized likelihood ratio test (GLRT), derived under the hypothesis of deterministic target amplitudes. While this conditional model is ubiquitous, we investigate here the equivalent GLR approach for an unconditional model where the target amplitudes are treated as Gaussian random variables at the design of the detector. The GLRT is derived which is shown to be the product of Kelly's GLRT and a corrective, data dependent, term. Numerical simulations are provided to compare the two approaches.