Comparison of suboptimal iterative space-time receivers

At high spectral efficiencies, optimal decoding of space time block coded (STBC) multiple input multiple output (MIMO) systems has a critical complexity unless the space-time code structure offers simple optimal decoding, as is the case for space time orthogonal block codes (STOBC). Indeed, the complexity of true maximum likelihood (ML) decoding or a posteriori probability (APP) decoding of STBC increase exponentially with the spectral efficiency. As a soft input soft output (SISO) decoding scheme, APP space-time decoding is of special interest since it allows the space-time receiver to efficiently work with a channel decoder. Besides, it can advantageously be used in an iterative manner with a SISO channel decoder to approach the optimal joint receiver performance. List sphere decoding (LSD) provides a good approximation of APP decoding with reduced complexity and can therefore be considered as a performance reference. This paper presents a new suboptimal SISO space-time receiver based on MMSE using priors, which has significantly lower complexity than LSD but turns out to perform similarly when used in an iterative receiver. The system under consideration is first presented, followed by a detailed description of the new proposed MMSE based receiver and a brief description of the reference LSD receiver. These two schemes are then evaluated in iterative receivers in terms of performance and complexity.