| 采用核函数的多用户检测算法 |
| |
| 引用本文： | 杨涛,谢剑英. 采用核函数的多用户检测算法[J]. 上海交通大学学报, 2004, 38(Z1): 49-52 |
| |
| 作者姓名： | 杨涛 谢剑英 |
| |
| 作者单位： | 上海交通大学,电子信息与电气工程学院,上海,200030 |
| |
| 摘 要： | 提出了一种基于核函数的多用户检测(MUD)方案,与常规的支持向量机(SVM)学习算法不同的是,判别输出函数中的支持向量采用一种稀疏核逼近方法获取,而其对应系数则由输入采样协方差矩阵的广义特征向量构成,整个算法避免了常规的二次规划(QP)求解过程.仿真结果表明,采用核函数算法的检测性能与SVM检测性能接近,但在较大规模样本集下可有效减小计算量.
|
| 关 键 词： | 核函数 支持向量机 多用户检测 |
| 文章编号： | 1006-2467(2004)S1-0049-04 |
| 修稿时间： | 2003-11-30 |
Multiuser Detection Algorithm Based on Kernel Function |
| |
| YANG Tao,XIE Jian-ying. Multiuser Detection Algorithm Based on Kernel Function[J]. Journal of Shanghai Jiaotong University, 2004, 38(Z1): 49-52 |
| |
| Authors: | YANG Tao XIE Jian-ying |
| |
| Abstract: | A multiuser detection (MUD) algorithm based on kernel function was proposed. Differing from that of support vector machine (SVM), the support vector (SV) in our scheme is obtained through a kernel sparsity approximation algorithm. The corresponding SV coefficient is constructed by generalized eigenvector of input sample covariance matrix. Thus the conventional costly quadratic programming (QP) computation is avoided. The simulation results show that the scheme has a comparable performance while with a reduced computation complexity. |
| |
| Keywords: | kernel function support vector machine multiuser detection |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
