Noise protection of existing radio lines with noise-shaped signals and digital types of modulation was studied. Analysis has shown that the use of such signals in conditions of the radio-electronic conflict does not permit to provide necessary level of noise immunity and transmission security of radio communication lines. It was explained by presence of cyclo-conditionality of the carrier oscillation in signals with digital modulation types. Such properties simplify detection and search of signals by means of spectral correlation methods of modern hostile means of electronic surveillance. To solve this problem, the use of nonstationary signal structures with variable central frequency and spectral density of power was proposed. A procedure of forming such signal structures by application of the Gram-Schmidt orthogonalization procedure to the ensemble of multicomponent LFM signals with controlled spectral characteristics was developed. It was proposed to estimate various signal structures of multicomponent signal by means of phase portraits of summed signals depending on the scaling factor value. This factor’s boundary values at which complexity of the multicomponent signal structure is ensured and degeneration of the process into classical LFM is prevented were established. Change of probability of a symbol error in a channel with the use of multicomponent orthogonal signal structures was studied depending on the signal/noise ratio. This makes it possible to estimate potential noise immunity of the radio line provided that the signal/noise ratio is determined by energy indicators of the radio channel and the spectral density of the noise of natural origin. Structural security of the developed signal structures was estimated by means of an energy detector and a cyclo-stationarity detector. It was established that in the case of energy detection, nonstationary signals, and signals of any other type of modulation are equivalent. However, probability of detecting nonstationary signal structures decreased 2–2.5 times compared to other types of signal modulation when using the cyclo-stationarity detector