针对经典盲均衡算法收敛速度较慢和稳态误差较大的问题,提出了一种基于变步长恒模算法(Constant Modulus Algorithm,CMA)和判决引导的最小均方(Decision Directed Least Mean Square,DD-LMS)算法的双模式切换盲均衡算法.在算法收敛初期采用 CMA 算法,以确保算法可以较快收敛.在收敛之后切换至 DD-LMS算法,以进一步降低稳态误差.通过设定阈值来切换算法,取相邻多次迭代误差的平均值作为算法的切换值,以确保算法切换时机的合理性.另外,引入 Softsign变步长函数并加入 3 个参数对该函数进行改进,使得 Softsign变步长函数可以依据不同信道环境设定最佳参数,同时提高算法的收敛速度.仿真结果表明,在卫星通用信道条件下,所提算法的收敛迭代次数约为 1 000 次,稳态误差为-12 dB,在信噪比为 15dB时,误码率为 1×10-6.与相关算法对比,所提算法的收敛速度较高,误码率和稳态误差较低.
A dual mode switching blind equalization algorithm which includes the constant modulus algorithm(CMA)and decision directed least mean square(DD-LMS)algorithm is proposed for the problem of slow convergence and large steady-state error in classical blind equalization algorithms.In the early stage of the convergence,CMA is used to ensure fast convergence.After the conver-gence,it switches to the DD-LMS algorithm to further reduce steady-state errors.By setting a threshold to switch algorithms,the average of adjacent iteration errors is taken as the algorithm switching value,ensuring the rationality of the switching timing.In addition,introducing the Soft-sign variable step function and adding three parameters to improve the function allows the Softsign variable step function to set the optimal parameters according to different channel environments,while also improving the convergence speed of the algorithm.The simulation results show that the proposed algorithm has a convergence iteration of about 1 000 times and a steady-state error of-12 dB in a satellite universal channel.When the signal-to-noise ratio is 15 dB,the error rate is 1×10-6.Compared with the relevant algorithms,the proposed algorithm has a higher convergence speed,lower error rate and steady-state error.
NETL
NSTL
维普
万方数据
