Reduced-Complexity Near-Maximum-Likelihood Detection for Decision Feedback Assisted Space-Time Equalization

Abstract

A novel decision-feedback (DF) aided reduced complexity maximum likelihood (ML) space-time equalizer (STE) designed for a single-carrier system is introduced. Two different methods of incorporating DF into the recursive tree search based receiver structure are proposed for allowing detection at a moderate computational cost. Additionally, a further complexity reduction scheme is proposed, which exploits the specific characteristics of both the wide-band channel and the proposed DF-STE. In comparison to the DF-STE not benefiting from this complexity reduction, the proposed detector is capable of reducing the complexity by several orders of magnitude. More quantitatively, for the specific rank-deficient system considered, which detected the signal transmitted from four transmit antennas with the aid of two receive antennas, the complexity might be reduced by a factor of 100 at low Signal-to-noise ratios (SNRs) without noticeable performance degradation. By contrast, at higher SNRs a complexity reduction of a factor of 10 might be achieved, depending on the tolerable performance degradation.