α噪声下的改进频域多模盲均衡算法

An Improved Frequency-domain Multimodulus Blind Equalization Agorithm under α Noise

针对卫星通信信号均衡中常数模算法在α稳定分布噪声中性能退化,及均衡多模信号时收敛速度较慢、稳态误差大等问题,提出了一种改进的频域多模盲均衡算法。该算法将模值变换、快速傅里叶变换及重叠保留法融入到分数低阶统计量的盲均衡算法中,利用分数低阶统计量对α稳定分布噪声进行抑制,根据多模信号星座图的分布特点,模值变换将多模信号的多个幅度模值变换成单一幅度模值,快速傅里叶变换及重叠保留法减少了传统盲均衡算法的计算量,同时对均衡器输入信号进行正交小波变换,减小输入信号之间的自相关性。仿真结果表明,该算法的稳态误差较小且收敛速度快,有效地提高了信号质量。

When constant modulus algorithm is used to equalize muhimodulus signals, there are some defects of the slow convergence rate and big steady mean square error, and its periormance degenerates in α-stable distribution noise. In order to overcome these disadvantages, an improved frequency-domain muhinlodulus blind equalization algorithm is proposed. In the proposed algorithm, modulus transform, fast Fourier transform and overlap-save law are combined into blind equalization algorithm based on fractional lower order statistics, fractional lower order statistics is used to restrain α-stable distribution noise, according to the feature of muhimodulus signal constellations, modulus transform changes the muhimodulus of muhimodulus signal to single modulus, the multimodulus signal is equalized effectively. The algorithm＇s computational loads is greatly reduced by using fast Fourier transform and overlap-save law, orthogonal wavelet transform is used to reduce the autocorrelation of the equalizer input signals. The theoretical analyses and the computer simulations in satellite channels indicate that the proposed algorithm can obtain fast convergence rate and the small steady mean square error. So the proposed algorithm has important reference value for the satellite communications.