结合最小均方误差的改进球形译码检测算法

在多输入多输出(MIMO)系统的信号检测算法中，球形译码算法的检测性能最接近最大似然算法，但传统球形译码算法运算复杂度较高。为降低球形译码算法复杂度，提出一种新型的球形译码检测算法。新算法由改进的快速球形译码算法与最小均方误差算法相结合而成。改进的快速球形译码算法通过在球形半径收缩时乘上一个常量参数来提高半径收缩速度，减少算法搜索的信号点数，从而达到降低复杂度的目的。最小均方误差算法则能够通过减小噪声对接收信号的干扰来降低因搜索噪声点而产生的复杂度。将最小均方误差算法的信道矩阵应用在改进的快速球形译码算法中，将两种算法有效地结合，能够进一步降低算法复杂度。仿真结果表明，当信噪比(SNR)低于10 dB时，新算法相比于原始球形译码算法，检测性能平均提高了9%左右。

Improved sphere decoding detection algorithm combined with minimum mean square error

Among all of the signal detection algorithms in multiple-input multiple-output systems,the capability of sphere decoding algorithm is most close to the capability of maximum-likelihood algorithm.But the calculation complexity of the sphere decoding algorithm is still very high.To decrease the calculation complexity of sphere algorithm,a new sphere decoding algorithm was proposed.The new algorithm was combined by an improved fast sphere decoding algorithm and the Minimum Mean Square Error(MMSE) algorithm.The improved fast sphere decoding algorithm can increase the decreasing rate of sphere radius via multiplying the contraction process of sphere radius by a constant parameter,so that it can reduce the number of signal points in search process to decrease calculation complexity.Meanwhile,the MMSE algorithm can reduce the interference that caused by noise,so that it can decrease the calculation complexity caused by the process of searching noise points.The channel matrix of the MMSE algorithm was applied to the improved fast sphere decoding algorithm,so these two algorithms can be combined with each other efficiently,and the combined algorithm can further reduce the calculation complexity.The simulation results show that,when Signal-to-Noise Ratio(SNR) is less than 10dB,the proposed algorithm improves average performance by 9% compared with original sphere decoding algorithm.