Two-stage-ranking assisted symbol detection for massive MIMO systems

摘要

The paper presents a maximum likelihood symbol detection scheme with the assistance of a two-stage-ranking mechanism for massive MIMO systems. The proposed scheme first computes the QR decomposition to the channel matrix in order to obtain them corresponding unitary and upper triangular matrices. Next, the ML mechanism is conducted to recursively estimate the transmitted symbols layer by layer. In particular, to enhance the performance, the proposed scheme conducts the first stage ranking mechanism to sort the channel matrix column vectors' norms and re-permutes the column vectors from the right to the left sides in the decreasing order of the vectors' norms before the QR decomposition. In addition, to further enhance the performance of the ML mechanism, a searching radius mechanism and the second ranking stage are embedded into the proposed scheme for each layer's symbol detection. Finally, the results of simulation and complexity analysis show that the proposed scheme outperforms existing methods, especially in high interference scenarios.