一种改进的QR分解检测算法

A modified algorithm of QR decomposition

结合排序和循环迭代的思想,提出了一种基于PIC预处理的循环迭代SQRD检测算法。该改进算法对排序后的信道矩阵进行循环迭代检测,在每次循环迭代检测过程中,利用PIC思想对检测数据进行更新。在得到所有层的判决值后,丢弃其它层的检测值,只保存最后检测层的判决值,并作为下次循环检测的最先检测值。算法同时克服了传统QR分解检测算法的误差扩散和最先检测层分集效益低的缺点。仿真结果表明,改进算法的检测性能要优于SQRD检测和循环迭代检测算法。

Combining the thought ol sorting and loop iteration,this paper proposes a cyclic iteration SQRD detection algorithm based on the PIC preprocessing. The improved method implements loop iteration detection in sorted channel matrix, using PIC to update decision value in every iteration detection process. After getting decision values ol all layers, it only saves the decision value in the linal detection layer and discards values in other layers, then using retained values as first detection value in next loop detection. The new method overcomes the error diffusion and low diversity benefit in the first detection layer ol traditional QR decomposition algorithm. The simulation results show that the detection performance of improved algorithm is superior to the SQRD detection and iterative detection algorithm.