RANK-BASED DETECTION OF WEAK RANDOM SIGNALS IN A MULTIPLICATIVE NOISEMODEL

Multiplicative noise is known to be useful in modeling some environmen t, which is difficult to describe by additive noise model. In this pap er, nonparametric detection of weak random signals in multiplicative n oise is considered. The locally optimum detector based on signs and ra nks of observations is derived for good weak-signal detection performa nce under any noise probability density function. The detector has sim ilarities to the locally optimum detector for random signals in multip licative noise. It is shown that the nonparametric detector asymptotic ally has almost the same performance as the locally optimum detector. (C) 1997 Published by Elsevier Science B.V.