基于矢量消息传递的迭代均衡

在迭代均衡中，通过计算后验概率密度求解输入信息可以被视为一个广义线性回归问题。为解决此问题，采用了一种名为广义矢量消息传递的新算法。该算法与之前的广义消息传递算法相比，可以适用于包含任意分布的输出。使用该算法的检测器中，包含一个内部软均衡器和软分块交织器，原理与软均衡器和软译码器的消息传递类似，其按照迭代算法逐块交换外信息，从而改善均衡器的性能。通过对该算法的性能分析，提出增加半自适应的阻尼系数来保证内部软均衡器和交织器的独立性。对改进后的算法进行性能分析，结果表明其计算复杂度和低信噪比条件下性能均优于逐符号的迭代均衡以及广义消息传递算法的迭代均衡。

Iterative equalization based on generalized vector approximate message passing

As performing interative equalization, jointly recovering the vector and the matrix from component nonlinear measurements can be considered as a generalized bilinear recovery problem.To address this problem,a novel algorithm called the bilinear adaptive generalized vector approximate message passing soft frequency-domain equalizer(BAd-GVAMP-SFDE) is proposed, which extends the generalized approximate message passing(GAMP) algorithm to incorporate arbitrary distributions on the output transform.The dector adopting the algorithm consists of an inner soft equalizer(ISE) and an inner soft slicer(ISS), which iteratively exchanges block-wise extrinsic information to improve the equalization performance.The performance analysis is provided for the GVAMP-SFDE and it leads to an improved semi-adaptive damping(SAD) scheme for maintaining the independence between the ISS and ISE.Performance analysis of SAD-VAMP-SFDE shows that it improves the mean square error(MSE) evolution curves compared with the original VAMP-SFDE.The SAD-VAMP-SFDE-based Turbo equalization outperforms those using the SI-BSIC-SFDE or GAMP-SFDE in low signal-to-noise ratio condition.