一种格缩减辅助MIMO检测的组合量化误差校正方法  

Quantization Error Correction Scheme for Lattice-reduction Aided MIMO Detection

作　　者：郑俊松[1] 周晓方[1] 

机构地区：[1]复旦大学专用集成电路与系统国家重点实验室,上海201203

出　　处：《小型微型计算机系统》2013年第8期1926-1929,共4页Journal of Chinese Computer Systems

基　　金：国家重大专项项目(2011ZX03003-003-03)资助;复旦大学专用集成电路与系统国家重点实验室自主项目(11MS003)资助

摘　　要：格缩减技术(LR)可以用于提升多入多出系统(MIMO)线性和非线性检测的性能.采用该方法后会引起检测信号星座图空间的畸变,导致变化后信号取值的非均匀分布和量化错误的易扩散性,会阻碍检测性能的提高.为了进一步提升检测性能,提出奇偶量化的组合量化误差校正方法.仿真结果显示,加入该方法的格缩减辅助检测的性能得到了明显的提升,而且可以很好的逼近最大似然检测(ML)的性能.和目前已知的其它同类量化误差校正方法相比,在实现相同的检测性能提升时,本文提出的组合量化误差校正方法增加的候选矢量减少了一半,即增加的运算复杂度最低.Lattice-Reduction technology has been proven that can enhance the performance of both linear and nonlinear MIMO detec- tion algorithms. However, the signal transformation of Lattice-Reduction aided detection will introduce the distortion of the constella- tion, and cause the diffusion of quantization error, which may finally increase the bit error rate. This paper proposes an improved par- ity-based quantization error correction scheme. Simulation results show that the performance of lattice reduction aided MIMO detection based on the proposed quantization scheme has a significant improvement, and is very close to ML. Compared to other published quantization error correction schemes, this scheme can achieve the same performance at the cost of a half number of candidate sym- bols, i.e. with the lowest additional complexity.

关 键 词：MIMO系统 格缩减 奇偶量化 量化误差校正 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]