摘要
常模盲均衡算法采用固定步长时存在收敛速度和稳态误差之间的矛盾,解决这一矛盾的有效方法是用变步长代替固定步长。在分析了常模算法误差特性的基础上,建立了步长因子和误差信号之间的一种非线性函数关系,分析了函数中参数对步长因子的控制,较好地满足了常模算法迭代过程对步长因子的要求,得到一种新的变步长常模算法。新算法能够在加快收敛速度的同时减小稳态误差,且具有较强的信道跟踪能力和抗干扰能力。理论分析和计算机仿真证实:新算法性能较好,具有很好的实用性。
There will be conflict between the convergent rate and the steady - state is error when the constant step - size is adopted in the constant modulus blind equalization algorithm. An effective way for solving this conflict is to employ variable step - size in stead of constant step - size. In this paper a nonlinear function relationship between the step - size and the error signal is established based on the analysis of the error signal character of the constant modulus algorithm, and the parameters in the function for controlling the step - size are analysed. The requirement of constant modulus algorithm on the step - size is greatly satisfied and the new algorithm can speed up the convergence and keep less steady - state error simultaneously, and it could do better at following the channel changes and resisting to noise disturber. Theoretical analysis and computer simulations show that the performance of the new algorithm is better than that of the conventional ones, thus is of practical importance.
出处
《计算机仿真》
CSCD
2008年第3期107-110,共4页
Computer Simulation
基金
总装备部探索研究项目基金资助(6130320)
关键词
变步长
非线性变换
常模算法
盲均衡
Variable step - size
Nonlinear transform
Constant modulus algorithm （CMA）
Blind equalization
