基于瑞利分布的变步长常模盲均衡算法

Rayleigh Distribution Based A Novel Variable Step-size Constant Modulus Algorithm

针对常模算法(CMA)及基于剩余误差非线性变换变步长常模算法(REVCMA)收敛速度慢的缺点,提出了以误差函数为随机变量、以瑞利分布为变步长,构造了基于瑞利分布的变步长常模盲均衡算法(RDVCMA)。该算法的变步长有两个参数,通过调整该参数,可以加快收敛速度。分别用4QAM及4PSK信号,通过典型电话信道,对算法进行了仿真研究。结果表明,在同样的稳态误差条件下,新算法具有比CMA及REVCMA更快的收敛速度。

Aiming at the slow convergence rates of Constant Modulus Algorithm （CMA） and Residual Error based Variable step-size CMA （REVCMA）, a novel Rayleigh Distribution based Variable step-size CMA （RDVCMA） is proposed. In this proposed algorithm, the Rayleigh distribution, which is a function of error and has two parameters, is used as variable sep-size. The convergence rates of the proposed algorithm can be improved by regulating its two parameters. Simulation tests with typical telephone channel using 4QAM and 4PSK are carried out respectively. The simulation results have indicated that the convergence rates of the proposed algorithm are faster than those of the CMA and the REVCMA under condition of the same steady state error.