The Modified Parallelizable Iteration Algorithm for Signal Detection in Massive MIMO Systems

摘要

As one of linear signal detection algorithms in massive multiple-input-multiple-output (MIMO) system, minimum-mean-square-error (MMSE) algorithm can reduce the computational complexity at the expense of a small loss in detection accuracy. However, the complicated matrix inversion involved in MMSE algorithm leads to low parallelism. The paper proposes a novel algorithm named modified parallelizable iteration (MPI) algorithm. MPI algorithm approximately calculates the matrix inversion in an iterative way, where the initial solution by use of the characteristic of the MMSE filtering matrix. Simulation results show that MPI algorithm can achieve the near-optimal performance and high parallelism. Meanwhile, the computational complexity is further reduced one order of magnitude.