The article shows that the solution to the problem of increasing the noise immunity (noise immunity and secrecy of functioning) of the ICS can be achieved using systems of nonlinear signals with improved ensemble, structural and correlation properties. Two classes of nonlinear complex discrete signals are considered: characteristic discrete signals (CDS) and cryptographic signals (CS). Methods for the synthesis of these signals are presented. The paper gives a statistical simulation model for studying the noise immunity of various classes of signals in the Gaussian channel. Using this model, estimates of the dependence of the error probability on the signal-to-noise ratio were obtained for various classes of signals, namely: CDS, KS and standard BPSK AFM-16 signals. It is shown that for the signal-to-noise ratio – 10 the error probability for the CDR is 4.6875e-06, for the CS is 3.515625e-06, and for the AFM-16 is 0.002025. Thus, the use of nonlinear complex discrete signals, in particular, CDS and KS, can significantly increase the noise immunity of signal reception in modern ICS. At the same time, taking into account the improved ensemble and structural properties of these nonlinear signals, it is possible to improve significantly the indicators of crypto- and imitation security of the systems functioning.