On the performance of digital detectors with dependent samples

The asymptotic performances of three digital detectors operating with dependent samples are examined. Analytic expressions for the performances when the number of dependent samples M is large are obtained. Examples are also given which indicate how the performances vary with M. It is shown that the performances of hard-limiting detectors may be improved by processing more dependent samples. The amount of the hard-limiting loss that is recoverable depends upon the detector prefilter. A simple expression is derived which gives the irretrievable hard-limiting loss as a function of the noise autocorrelation,.>