Reduced-dimension MAP turbo-BLAST detection

Abstract

Absfruct-The BeU-Labs Layered Space-time (BLAST) architecture is a simple and efficient multi-antenna coding S~N C~U I V that can achieve high-spectral efficiency [l]. Many BLAST detectors require more receiver antennas than transmitter antennas. We propose a novel turbo-prncessing BLAST detector based on a group detection strategy that can operate in system with fewer receiver antennas than transmitter antennas. A maximum a pnsterinri (MAP) decision is made using a group of transmitted symbols and the remaining signal contribution is treated as interference. The interference is characterized as non-zero mean colored noise source that is whitened before a decision is made.

SUMIS: Near-Optimal Soft-In Soft-Out MIMO Detection With Low and Fixed Complexity

“…We propose a new method that is inspired by the ideas in [11]- [14] of partitioning the original problem into smaller subproblems. As in the PM and RDMAP methods, we perform marginalization over a few of the bits when computing the LLR values.…”

“…The core idea behind SIS germinated in [22] in a different context than MIMO detection, but the use of SIS as a constituent of our proposed algorithm is mainly inspired by the work in [22]- [25]. The main differences between the SIS procedure used in our algorithm and that in [23]- [25] are: (i) we allow for the signal and interference subspaces to have varying dimensionality; (ii) we perform the SIS in a MIMO setting internally without the need for a priori information from the decoder, as opposed to [11], [23]; and (iii) we do not iterate the internal LLR values more than once, nor do we ignore the correlation between the interfering terms over the different receive antennas as in [24], [25]. We refer to our method as subspace marginalization with interference suppression (SUMIS).…”

“…and due to the assumption on S being BPSK, we can perform the marginalization in (11) in the LLR domain as (after inserting y in place ofȳ in (12))…”

“…III for higher-order constellations, this stage is performed symbol-wise; hence we use here again the index k. Note that for BPSK constellations, the procedure in [11] is equivalent to stage I here, i.e., the method of [11] is a special case of SUMIS when stage II is disabled.…”

“…These methods are more desirable from an implementation point of view, as they eliminate the need for data buffers and over-dimensioning (for the worst-case) of the hardware. Examples of such detectors are the reduced dimension maximum a posteriori (RDMAP) method [11], the partial marginalization (PM) method [12], and the fixed-complexity SD (FCSD) aided max-log method [13]. These fixed-complexity detectors provide a simple and well-defined tradeoff between computational complexity and performance, they have a fixed and fully predictable run-time, and they are highly parallelizable.…”