Performance evaluation of V-BLAST detection algorithms and a novel segmented detection algorithm

Several detection algorithms for the vertical Bell Laboratories layered space time (V-BLAST) system are compared. We propose the concept of segmented detection on the observation that the overall performance of decision feedback equalization (DFE) is limited by the performance of the first detected subchannel. We perform maximum-likelihood (ML) detection for the first several subchannels and use the DFE procedure to detect the remaining subchannels. Also, we propose a novel segmented detection algorithm based on the minimum mean square error (MMSE) criterion. Compared with DFEs, segmented detection improves the detection performance with little increase of complexity. Computer simulation verifies it.