摘要
针对时变信道环境下的盲源分离问题,在现有实时盲分离算法的基础上,提出了两种适用于时变信道环境的实时盲源分离算法——基于优选函数EASI盲分离算法(EASI-function)和基于选优函数的峭度变步长盲分离算法(EASI-function-KVS)。EASI-function算法在信号分离的不同阶段采用不同的估计函数,从而使得算法在收敛速度和稳态性能两方面获得一个折中。EASI-function-KVS算法则在EASI-function的基础上,利用峭度变步长的思想,自适应地调整迭代步长大小,进一步改善分离算法性能。仿真结果表明,两种算法能有效地跟踪信道变化,并且在性能方面比传统的EASI算法要好。
Aiming at the problem of blind source separation （BSS） in time varying channel, two real time BSS algorithms are presented： equivariant adaptive source separation via independence （EASI） BSS algorithm based on optimal selective function （EASI-function） and BSS based on optimal selective function algorithm with variable step-size based on kurtosis （EASI-function-KVS）. EASI-function algorithm adopts the different estimate function in different stage to get the balance between divergence speed and steady-state performance. And the EASI-function-KVS algorithm changes the step-size adaptively based on kurtosis thought so as to improve the performance of algorithm. Simulation results show that the algorithms in the paper can trace the channel changes effectively and the performance is superior to that of conventional EASI algorithm.
出处
《电子科技大学学报》
EI
CAS
CSCD
北大核心
2012年第4期512-515,共4页
Journal of University of Electronic Science and Technology of China
基金
高等学校学科创新引智计划(B08038)
中央高校基本科研业务费专项资金(721024669)
关键词
算法
盲源分离
信号处理
时变信道
变步长
algorithm
blind source separation
signal processing
time-varying channel, variable step-size
