A Pairwise Ordered Successive Interference Cancellation Algorithm for VBLAST

The classic detection algorithm of VBLAST, OSIC (ordered successive interference cancellation), is a good compromise between performance and complexity comparing to the optimum ML (maximum likelihood) algorithm. Due to the error propagation, the detection errors in the previous stages would degrade the accuracy of the forthcoming stages´ detection and the overall performance of OSIC. We propose, in this paper, an POSIC (pairwise OSIC) algorithm aiming at reducing the errors in each detection stage and the numbers of the stages in the detection process to diminish the error propagation. Instead of detecting one substream in each stage, the POSIC algorithm using ML to jointly detect two strongest sub streams in each stage and then eliminate the contribution of the detected substreams from the received signals until all the transmitted substreams were detected