针对后非线性盲源分离中非线性参数估计中存在的问题,提出一种基于改进的自适应遗传算法的后非线性盲源分离方法.该方法给出一种新的适应度函数,利用适应度函数值反馈调节交叉概率和变异概率的选取,并将优先进化策略和模拟退火机制引入遗传算法中,再通过线性分离算法得到分离矩阵.仿真验证表明,该方法较传统方法具有更快的收敛速度和较高的分离精度.
Considering the deficiency of nonlinear parameter estimation in post-nonlinear mixtures BSS, we propose a method based on an improved GA. First, the value of a new fitness function is used to reflect on the choice of crossover and mutation probabilities. PE and SA are incorporated in GA. Finally, the linear separation algorithm is applied to estimate the separation matrix. The results indicate that the new method has better performance than conventional algorithms.
[1] Wang SZ, Yang XK, Liu P, et al. Nonlinear blind source separation based on compound kernel function[J]. Journal of System Simulation, 2010,22(1):1-4 (in Chinese). 王曙钊,杨晓阔,刘鹏,等.一种基于组合核函数的非线性盲源分离方法研究[J].系统仿真学报,2010,22(1):1-4.
[2] Song K, Wang Q. Approach to nonlinear blind source separation based on niche genetic algorithm //The 6th Proc of Int Systems Design and Applications. Jinan: Elsevier Science, 2006: 441-445.
[3] Wang F S, Li H W, Shen Y T. An overview on nonlinear blind source separation: theory and algorithms [J]. Signal Processing, 2005, 21(3):282-288 (in Chinese). 王法松,李宏伟,沈远彤.非线性盲源分离的原理及算法综述[J].信号处理, 2005, 21(3):282-288.
[4] Tao T, Huang G M, Zhao Z H. Survey on blind source separation algorithms for post-nonlinear mixtures [J]. Control and Decision, 2010, 25(8):1121-1127 (in Chinese). 陶涛,黄高明,赵治华.后非线性混叠信号盲源分离算法综述[J]. 控制与决策,2010,25(8): 1121-1127.
[5] Ge J K, Qiu Y H, Wu C M, et al. Summary of genetic algorithms research [J].Application Research of Computers, 2008, 25(10): 2911-2916 (in Chinese). 葛继科,邱玉辉,吴春明,等.遗传算法研究综述[J].计算机应用研究, 2008, 25(10): 2911-2916.
[6] Zhan Z L, Zhang Y. Extraction of a source signal whose kurtosis value lies in a specific range [J].Neural Computing (S0925-2312),2006,69(7-9):900-904.
[7] Bi W H, Ren H M, Wu Q B. A new elitist strategy in genetic algorithm[J].Journal of Zhejiang University: Science Edition, 2006, 33(1): 33-35 (in Chinese). 毕惟红,任红民,吴庆标.一种新的遗传算法最优保存策略[J]. 浙江大学学报:理学版, 2006, 33(1): 33-35.
[8] Chen H, Cui D W, Yan T S, et al. Race exponent based simulated annealing rank selection operator [J]. ACTA Electronical Sinica,2009, 37(3):586-589 (in Chinese). 陈皓,崔杜武,严太山,等.基于竞争指数的模拟退火排序选择算子[J].电子学报,2009,37(3):586-589.
[9] Tan Y, Wang J. Nonlinear blind source separation using higher order statistics and a genetic algorithm[J].IEEE Transactions on Evolutionary Computation (S1089-788X), 2001, 5(6): 600-612.
[10] Taleb A, Jutten C. Source separation in post-nonlinear mixtures [J]. IEEE Trans on Signal Processing, 1999, 47(10): 2807-2820.
[11] Yang J A, Wang L W, Zhuang Z Q. Research of nonlinear blind source separation algorithm using higher order statistics and quantum genetic algorithm[J]. Pattern Recognition and Artificial Intelligence, 2004, 17(1):41-46 (in Chinese). 杨俊安, 王伦文, 庄镇泉.基于高阶统计理论和量子遗传算法的非线性盲源分离算法究[J]. 模式识别与人工智能,2004,17(1):41-46.
