基于CP分解的MIMO-OFDM系统接收信号盲检测

Blind Detection of Received Signal Based on CP Decomposition in MIMO-OFDM System

多输入多输出-正交频分复用(MIMO-OFDM)无线通信系统中接收信号从空间、时间、频率的维度形成多因素的阵列信号,传统的矢量或者矩阵代数的建模方法在处理多因素信号问题上显得不足,无法利用多因素间的关系,而张量分析在解决多维阵列信号处理的问题上具有优势。针对MIMO无线通信系统,结合OFDM技术,研究了张量信号的建模及分解方法,并充分利用张量信号的分解唯一性提高无线接收信号的检测能力。提出了基于CP(CANDECOMP/PARAFAC)张量分解方法对未知信道状态(CSI)的MIMO-OFDM系统进行接收端的张量信号建模和盲检测,并通过仿真分析验证了模型的可行性。仿真结果表明,在接收天线数目大于发送天线数目且各径信道独立情况下,基于CP分解的接收信号盲检测算法在误码率为10-4时,随着接收天线数目增加,信噪比可获得约5 d B的增益。

The received signal of multiple input multiple output-orthogonal frequency division multiplexing（MIMO-OFDM） wireless communication system forms a multi-factor array signal from the view of space,time and frequency dimension. The traditional vector or matrix algebraic methods cannot perform well whenprocessing multi-factor signals and cannot make use of the relationship among factors. Tensor analysis hasan advantage in processing multi-dimension array signal. Considering the MIMO wireless communicationsystem with OFDM technique,this paper researches on the tensor modeling and decomposition methods andaims to increase the detection ability of wireless received signal by use of the uniqueness of tensor decom-position. The received tensor signal with unknown channel state information（CSI） in MIMO-OFDM systemis modelled and detected blindly based on the CANDECOMP/ PARAFAC（CP） decomposition method. Sim-ulation results verify the feasibility of the modeling. When the number of receiving antennas is larger thanthat of the transmitting antennas and the multi-path signals are independent,the blind detection algorithmbased on CP decomposition can obtain about 5 dB gain in terms of signal-to-noise ratio（SNR） for bit errorrate（BER） 10^-4 with the number of receiving antennas increasing.