基于近似消息传递的VLC非线性均衡器

VLC Nonlinear Equalizer Based on Approximate Message Passing

沃尔特拉非线性均衡器(VS-NPE)是解决可见光通信(VLC)信道失真问题的一种可行方案,但却具有内核参数数量大、计算较困难的缺点。针对上述问题,提出一种高效的VS-NPE内核参数计算方法。结合沃尔特拉级数特点,修正内核参数观测矩阵以降低其列间相关度;在近似消息传递算法的基础上,引入阻尼因子以快速追踪可靠的稀疏支撑集,进而采用最小二乘算法计算出有效的内核参数。仿真结果表明:相比传统的参数计算方法,所提方案在短训练样本条件下依然可取得-31.55dB的NMSE,且使用的内核数量最少。上述方法不仅可有效地降低VS-NPE内核参数计算复杂度,且能取得优异的计算精度。

Volterra-based nonlinear equalizer(VS-NPE) is a feasible solution to solve the channel distortion problem in visible light communication(VLC),however, it has a large number of kernel parameters that are very difficult to calculate. Aiming at this problem, an efficient VS-NPE kernel parameter calculation method is proposed. First, according to the characteristics of the Volterra series, the kernel parameter observation matrix was modified so as to reduce the correlation between its columns;then, based on the approximate message passing algorithm, a damping factor was introduced to quickly track the reliable sparse support set, and then the least squares algorithm was used to calculate the effective kernel parameters. The simulation results show that: compared with the traditional parameter calculation methods, the proposed scheme can still achieve-31.55 dB NMSE under the condition of short training samples, and the number of used kernels is the least. This method can not only effectively reduce the calculation complexity of the VS-NPE kernel parameters, but also achieve excellent calculation accuracy.