摘要
针对循环前缀采用特定序列(UW)的多输入多输出单载波频域均衡(MIMO-SCFDE)系统,提出了基于噪声预测的频域均衡器.在均衡器的输出端,根据数据估计噪声和UW估计噪声的相关特性,由UW噪声来预测和抵消数据估计中的噪声,从而改善了系统性能.根据迫零(ZF)和最小均方误差(MMSE)准则,分别推导了ZF和MMSE噪声预测均衡器.理论分析说明,与传统的MIMO频域均衡器相比,ZF噪声预测均衡器的输出端可以获得更低的噪声功率,MMSE噪声预测均衡器输出端的均方误差更小.实验仿真表明,噪声预测的频域均衡方法与传统方法相比可以获得明显的性能增益.
In this paper,a noise predictive equalization scheme is proposed for single-cartier multiple-input and multiple-output frequency domain equalization （MIMO-SCFDE） system, which utilizes a unique-word （UW） as cyclic prefix. The noise in data estimate is predicted by exploitation of the deterministic characteristic of UW and the correlation of the noise at the output of frequency domain equalizers. The predictive results are used to cancel the noises contained in the estimation of the data and enhance the system performance.Based on zero-forcing （ZF） and minimum mean square error （MMSE） criteria,ZF and MMSE noise predictive equalizers are derived respectively. Theoretical analyses show that both of the proposed techniques outperform the conventional frequency domain equalizers. Simulation results have confirmed the significant performance gain they could achieve.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2009年第1期43-47,共5页
Acta Electronica Sinica
基金
国家自然科学基金(No.60772100)
上海市重大科技攻关项目(No.07DZ15006)
关键词
MIMO
频域均衡
噪声预测
multiple-input and multiple-output （MIMO）
frequency domain equalization
noise prediction
