Fast maximum likelihood decoding of quasi-orthogonal codes

In this paper, we present a new fast maximum likelihood (ML) detection algorithm for quasi-orthogonal space-time block codes ( QO-STBC). Observed in a transformed signal space, the QO-STBC encoder embeds information into a group of orthogonal subspaces of rank two. We propose a fast technique that separates information embedded in each rank-two subspace by a simple linear processing, and decodes the separated information independently. For a multiple-input and multiple-output (MIMO) system, decoding algorithm decomposes the original N × N system into N/2 parallel 2 × 2 MIMO sub-systems without information loss followed by MRC. It then detects symbols contained in each 2 × 2 MIMO sub-system using a sphere decoder of reduced computational complexity.