No abstract

No abstract

Classical ML decoders for multiple input multiple output (MIMO) systems like the sphere decoder, the SchnorrEuchner algorithm, the Fano and the stack decoders suffer from high complexity for high number of antennas and large constellation sizes. In this paper, we propose the use of parallel processing for stack decoding, to decode signals transmitted on linear MIMO channels to reduce time consumption of hardware architecture. It will be shown that the parallel stack decoder allows a 50% less of run time compared to the classical stack decoder.

In this paper, linear precoding for non-orthogonal space time block codes (STBC) is investigated. A theoretical model of spatial correlation with a Laplacian distribution of AOA is first derived. The design of the precoder is based on the choice of the codeword error matrix according to a criterion. We propose here a new criterion based on the system outage probability to select the suitable codeword error matrix allowing to move rapidly from one diversity order to the next. Codeword selection points out the importance of the determinant and the eigenvalues of the error matrices. The proposed method is applied to the non-orthogonal optimal STBC : 2 × 2 Golden Code and 4 × 4 Perfect Code.

No abstract