基于最小四阶矩准则的RLS空时多用户检测算法

为了解决常用高阶准则存在局部极小点问题。提出了一种约束最小四阶矩（LMF）盲多用户信号检测准则。它是最小均方准则（LMS）的一种高阶改进形式，通过Hessian矩阵对LMF目标函数的凸性分析,证明这种检测准则拥有全局收敛的特性,将此检测准则与最小二乘迭代方法相结合，在空时多径CDMA系统中构造出自适应RLS-LMF多址检测算法,继承了前者的稳态解相关性能和后者的计算有效性，在多径CDMA信道和直扩超宽带（DS-UwB）空时信遒下的仿真实验表明，该算法在保证高收敛速度的同时，拥有优异的稳态检测性能。

A space-time multiuser detection based on the RLS-LMF algorithm

A code-constrained least mean fourth(LMF) criterion is presented for blind multiuser detection.It is an improved higher-order form of the least mean square(LMS) criterion.Other than the commonly used existing high-order criterion local minima,the LMF is proved to be globally convergent.The Hessian matrix of the objective function is derived and analyzed to prove the global convexity of the LMF.In multipath CDMA channels,the LMF criterion is combined with the recursive least-square method and the RLS-LMF algorithm is further developed to construct a space-time receiver.Simulations show that this novel algorithm outperforms existing techniques in CDMA and ultra-wideband channels.