The transition-based adaptive filter: a new technique to suppress digital cochannel interference

Abstract

Most of the techniques proposed for interference suppression in digital communication systems are based on the assumption that the signal to be rejected has characteristics different from that of the signal of interest (SOI). Therefore, when the SOI and the signal not of interest (SNOI) have similar characteristics, the performance of these techniques degrade. This paper proposes a new technique to separate cochannel signals that have similar characteristics by exploiting the fact that different transmissions will have different data clock phases, and hence it is possible to separate the data transitions of the SOI from that of the SNOI. This principle is used to modify the conventional adaptive filter to reject interfering signals that may have similar characteristics. The proposed technique is used to modify the operation of the conventional least mean square (LMS) adaptive filter, although the technique can be readily extended to other adaptive algorithms and architectures. Simulation results are presented comparing the performance of the proposed technique with the conventional adaptive filter scheme.