基于序列检测的块判决辅助Turbo均衡器

该文针对高斯分布假设条件下逐符号检测均衡器的不足，提出了一种基于序列检测的块判决辅助Turbo均衡器(Turbo-BDAE)。新算法利用译码全信息估计发送序列，提高了序列估计的可靠性，引入软判决可以降低硬判决引起的错误扩散，同时结合序列检测为均衡器输出的后验概率(APP)估计提供了更可靠的信息。仿真表明，该均衡器在保持相近复杂度的同时，进一步提升了均衡系统性能，优于逐符号LMMSE均衡器和硬反馈BDFE。

Sequence-based Block Decision-Aided Equalizer for Turbo Equalization

A sequenced-based Block Decision-Aided Equalizer (BDAE) for turbo equalization is proposed, which can overcome the deficiency of symbol-by-symbol detection equalizers that the A Posteriori Probability (APP) of each symbol is calculated individually based on a Gaussian distribution assumption of the equalizer output. The proposed equalizer used the full decoder’s information instead of the extrinsic information to improve the reliability of the estimated sequences. Replacing hard decision with soft decision reduces the effects of the error propagation. Meanwhile, combining soft decision with sequence-based detection provides more reliable information for the calculation of the APP. Simulation results show the new proposed equalizer leads to better performance compared to the Linear Minimum Mean Square Error (LMMSE) equalizers and the BDFE with hard feedback without sacrificing the computational complexity.