基于几何均值分解的可见光MIMO预编码优化设计

Optimization of precoder design for VLC-MIMO system based on GMD

在可见光通信(VLC,Visible Light Communication)系统中,受发送信号非负约束和信道相关性较强影响,预编码后的等效子信道增益差异较大,现有的预编码需要迭代搜索等复杂的功率和比特分配算法来适配不同子信道。针对此问题,利用非线性判决反馈均衡检测器,提出基于几何平均分解的线性预编码,将信道分解成多个增益相同的子信道来提高误码率性能,避免了传统预编码中复杂的功率和比特分配问题,又基于最小均方误差准则进行功率分配,进一步提高了系统性能。最后仿真实验分析结果表明,所提两种方案均明显优于传统预编码,误码率性能获得了约8 dB~12 dB增益。

In Visible Light Communication(VLC)systems,due to the influence of the non-negative constraint of the transmitted signal and the strong channel correlation,the equivalent subchannel gains after precoding are very different.The existing precoder requires complex power and bit allocation algorithms such as iterative search to adapt to different subchannels.Aiming at this problem,the nonlinear decision feedback equalization detector is used by this paper.Based on the geometric mean decomposition for linear precoder,the channel can be decomposed into multiple subchannel with the same gain to improve the bit error rate performance and the complexed power and bit allocation problems of the traditional precoder are avoided,and then based on minimum mean square error criterion for power distribution,the BER performance is further improved.The experimental results show that the proposed two schemes are significantly better than the traditional precoder,and achieve about 8 dB-12 dB performance gain in terms of bit error rate.