| 摘要: |
| 为了降低FastICA算法的计算复杂度，提出了一种基于多用户检测串行干扰抵消的新型独立分量分析算法MUD_FastICA。该算法结合了盲信号分离和多用户检测串行干扰抵消两种信号处理技术，利用减法和低维特征值分解来保证每次分离出不同独立分量和达到降低算法复杂度的目的。通过分析和仿真可以看出，所提算法在不影响分离性能的前提下，显著降低了算法的迭代次数和每次迭代的计算复杂度。在信噪比0 dB和4个源信号混合情况下，分离第二个信号的迭代次数和所需计算单元分别下降了14%和37%，分离第三个信号的迭代次数和所需计算单元分别下降了22%和58%，因此更加适合对实时性要求高的通信系统。 |
| 关键词: 多用户检测 独立分量分析 FastICA算法 串行干扰抵消 算法复杂度 |
| DOI： |
|
| 基金项目:国家自然科学基金资助项目(51275405) |
|
| An improved independent component analysis algorithm based on multiuser detection |
| MU Chang,YAO Jun-liang |
| () |
| Abstract: |
| A new independent component analysis （ICA） method named MUD_FastICA is proposed to reduce the computation complexity of FastICA algorithm. In the proposed method, multiuser detection(MUD) successive interference cancellation(MUD-SIC) is combined with one-unit FastICA to separate different independent components and reduce computation complexity. Analysis and simulation results show that, the MUD_FastICA can reduce the number of iteration and computation complexity obviously. Meanwhile,the separation performance is approximately the same with that of traditional FastICA. In the case of 0 dB SNR and four source signals, number of iterations and computing units to separate the second (third) signal are decreased by 14% (22%) and 37% (58%) respectively. Hence, it is more suitable for real-time communication system. |
| Key words: multiuser detection ICA FastICA algorithm successive interference cancellation computation complexity |
《电讯技术》编辑部 