Generalized signal-to-noise ratio and its estimation

A generalized realized signal-to-noise ratio (generalized RSNR) is defined for the situation where there are r, p-dimensional, real, or complex valued signals and a transmitted message is to be classified as either one of them or as pure white noise. Certain estimators of generalized RSNR are suggested under various loss functions. The closest estimator in the sense of Pitman is obtained. Finally, a lower confidence bound is provided.<
>