子空间约束的最小二乘常模算法

A Subspace Constrained LSCM Algorithm

常模算法(CMA)是一种性能优良的盲多用户检测算法。最小二乘常模算法(LSCMA)因其全局收敛性和稳定性而备受关注,但是在信噪比低时性能不理想。本文将最小二乘常模算法,紧缩近似投影子空间(PASTd)算法和奇异值分解(SVD)相结合,提出了一种子空间约束的最小二乘常模算法,称为SUB_LSCMA,其复杂度比已有的基于直接对接收信号自相关矩阵做特征值分解(ED)的LSCMA_SUB复杂度低。仿真结果表明这种算法的收敛速度、跟踪性能和误码性能和LSC- MA_SUB基本相同。

The constant modulus algorithm（CMA） is a blind algorithm. It can be applied in multiuser detection. The least squares constant modulus algorithm（LSCMA） is concerned because of its globally convergence and stability, but its performance is not superior when the SNR is low. In this paper we combine LSCMA ,Projection Approximate Subspace Tracking with deflation（PASTd） algorithm and Singular Value Decomposition （SVD）, and propose a subspace constrained LSCM algorithm（SUB_LSCMA）. Moreover, the computational complexity of the proposed SUB_LSCMA is lower than that of LSCMA SUB which is based on the eigenvalue decomposition （ED）of the received signal autocorrelation matrix. Simulation results show that the proposed SUB_LSCMA is similar to the LSCMA_SUB on the convergence rate, the tracking ability and the BER performance.