一种改进的变步长块恒模盲均衡算法

摘要

在块恒模盲均衡算法(BCMA)的基础上,引入可变步长因子,提出了一种新的变步长BCMA算法(VSS-BCMA).该算法的实时步长会随着块平均误差增加或减小而相应变大或变小,在算法初始阶段步长较大,而在趋于稳态时步长较小.实验仿真和分析表明,该算法与原BCMA算法相比,具有更快的收敛速度,更小的稳态误差,且信道均衡性能良好,易于实时分块并行处理.该算法比传统的变步长CMA算法运算效率更高,有利于并行运算,节省了时间,提高了处理器的实时处理速度和效能.%Based on BCMA (block constant modulus blind equalization algorithm) and by introducing variable step factor,a new VSS-BCMA (variable step-size BCMA) algorithm is proposed.The real-time step-size of the algorithm would increase or decrease with raising or dropping of the averaged block error.At the initial stage of the algorithm,the step-size is fairly large,while at the steady state relatively small.Simulated experiments and thoretical analysis indicate that the new algorithm,as compared with traditional BCMA algorithm,is of faster convergence speed,smaller steady-state error,with good performance of channel equalization and easy real-time parallel processing,and also as compared with traditional variable step-size CMA algorithm,the algorithm has an even higher operation efficiency,is easy for parallel operation,and thus could save time and improve real-time processing speed and efficiency of the processor.