FIELD: radio engineering, communication.
SUBSTANCE: disclosed method involves compressing a received signal in a compression filter in a restricted channel, comparing the compressed signal with a detection threshold and making a decision on signal detection if the compressed signal exceeds the threshold. The received signal is further compressed in a compression filter in a linear channel; the level of the compressed signal is compared with a linear channel threshold; a decision on detection of a k-th signal is made, where k is the index number of the signal compressed at a time tk and having a level Urestk in the restricted channel, that has not reached the threshold in the restricted channel if that signal in the linear channel has a level Ulink higher than the linear channel threshold and if in the interval tk±T, where T is the duration of the emitted signal, there is an i-th compressed signal, where i is the index number of the signal compressed at a time ti, having a level Ulini higher than the detection threshold in the linear channel, and the corresponding compressed signal detected in the restricted channel and having a level Uresti and the condition 
EFFECT: false alarm probability stabilisation while retaining the capacity to detect a weak signal during partial overlapping thereof with a stronger signal.
