For the new asymmetric paired carrier multiple access (APCMA) signal with the station signal bandwidth greater than the main station signal

on the basis of a low-complexity blind separation structure

an efficient blind separation algorithm based on the complementary symmetric filters is proposed. By constructing the complementary symmetric filters

the algorithm decomposes the mixed signal into two component signals in the time-frequency domain

one containing both the main station signal and the station signal

and the other containing only the station signal. Further

it separates the signal component containing the main station signal in the case of the same sample rate. Then

it synchronizes the two signal components under the same parameters

which ensures the additivity and integrity of the station signals in the two signal components after the separation. The experimental results show that compared with the original low-complexity separation algorithm

the proposed algorithm effectively improves the separation performance of the mixed signal. In addition

the complementary symmetric filters with information invariance proposed in this paper have a wide application prospect

which can be used in broadband multi-signal cancellation

channel estimation and spread spectrum detection under covert transmission.