Iterative interference cancellation and decoding for convolutional coded VBLAST systems

In this paper, we present an iterative interference cancellation (IIC) and decoding scheme that significantly improves the performance of coded VBLAST systems. We show that due to the residue interference caused by incorrect decisions from the previous iteration, the distribution of the total noise in the IIC process is non-Gaussian. Euclidean distance therefore cannot be used as the bit metric in soft decision Viterbi decoding. To preserve the Gaussian distribution of the total noise, we propose to use soft ing for VBLAST detection. We also propose a scheme employing soft decision Viterbi decoding with IIC (SDD-IIC) and demonstrate that it significantly improves both the coding gain and diversity gain of the VBLAST system.