An extension of the general method of obtaining an optimum system developed by the author

is given to include the case of nonlinear dependence of the observed function on signal parameters. The method affords effective determining of optimum systems designed for the detection and reproduction of signals in the presence of noise using various practically adequate criteria.