A Low Complexity Vector Perturbation Precoding with Projection Analysis Assisted Subspace Searching

In this paper, a novel low-complexity vector perturbation (VP) precoding is presented for multiple-input multiple-output (MIMO) systems with projection analysis (PA) assisted subspace searching method as VP-PA. It is verified with the projection analysis in this paper that if the perturbed transmit signal vector is in the subspaces spanned by the right singular vector with the least singular value of its precoding matrix, the transmit signal power is reduced to a shortest projection sum. Therefore, the conventional full searching space for VP method is significantly shrunk in the above subspaces to reduce the complexity. However, the best perturbation vector to fall in the above subspaces can't always be an integer perturbation vector, thus our approach involves searching integer perturbation vector in the shrunk subspaces to minimize the transmit power and reduce the complexity. The complexity and performance of the proposed VP-PA are analyzed and compared with conventional VP methods, which is proved to be a good choice for VP method considering both the performance and complexity.