高斯噪声下基于EM梯度算法的多用户检测的应用

Multi-user Detection Based on EM Gradient Algorithm in Gaussian Noise

EM算法是近些年来常用的求解缺失数据时的迭代算法,EM算法优点是简单易行性与普遍适用性。如果最大化不能用解析的方法实现而采用一种迭代优化的方法实现M步将产生嵌套循环计算,由此增加了计算的复杂度。基于此提出了EM梯度算法。在DS-CDMA系统中应用EM梯度算法进行检测,对比EM算法下的误码性能与收敛性能,实验结果显示EM梯度算法比标准EM算法具有相同的收敛速度。

The EM algorithm is commonly used for the missing data of iterative algorithm in recent years.Even if E-step is easy to implement,the calculation of M-step is also a great difficulty.In order to reduce the computational complexity of the M-step,we proposed the EM gradient algorithm.In MUD system within Gaussian noise,simulation numerical is carried out.The results show that the EM gradient algorithm has a good BER performance curve.It also has the same convergence speed as standard EM algorithm.