基于降维预处理独立分量分析的多用户检测

Multi-user detection based independent component analysis with preprocessing of dimension reduction

在由多径传播而产生的快衰落信道中,独立分量分析(ICA)算法产生性能恶化。为了消除这个缺陷,在进行ICA运算前,先对接收数据进行降维预处理。实验结果表明:在接收数据很少时,提出的多用户检测方法的性能大大优于一般的基于独立分量分析的多用户检测的性能,具有更好的实际适用性。

In the multi - path fast- fading channel, the independent component analysis （ICA） algorithm could not be convergent and the result would be deteriorative. In order to eliminate this disadvantage, dimension reduction is executed before ICA algorithm. The simulation results show that even in the case of small size of samples, the proposed method has better performance compared with multi - user detection based on ICA without dimension reduction. So the proposed method has much practicability.