摘要
针对常规的小波变换并不能保证对信号实现最优分解。以时频性能优良的Morlet小波为单一基函数,并扩大了基函数的参数空间,确定了满足小波容许条件的参数空间边界,提高了用其分解信号的灵活性。对实数编码遗传算法进行了改进,改进后的算法免除了独立的变异操作,提高了算法的搜索速度、收敛性和准确性。以互信息最小化为寻优的判据准则,给出了基于改进的遗传算法和改造后的Morlet小波的信号最优分解的具体步骤。对时变信号的仿真分析表明所提方法是有效和可行的。
The main function of wavelets transform is to decompose signals, but routine wavelets transform can not assure the optimal result of signal decomposition. Then a method of optimal decomposition for time-varying signal is present based on the morlet wavelet and improved genetic algorithm. In this method, morlet wavelet is reconstructed to enlarge the parameter space by changing the scal and delay factors to be variable, and the boundary of parameter space that satisfies the admissible condition of wavelet is confirmed. The minimum mutual information is selected as criterion when decomposing signal. As the method of optimization, real code genetic algorithm is improved to increase the convergence rate, success rate and anti-premature ability. The process of decomposition based on the improved genetic algorithm and reconstructed morlet wavelet is present, and the result of analysis shows that the method is effective and feasible.
出处
《辽宁工程技术大学学报（自然科学版）》
EI
CAS
北大核心
2007年第4期592-595,共4页
Journal of Liaoning Technical University (Natural Science)
基金
辽宁省教育厅攻关项目(2004D089)
关键词
小波变换
信号分解
遗传算法
互信息
寻优
wavelets transform： signal decomposition： genetic algorithm： mutual information： optimization
作者简介
李允公（1976-）,男，辽宁抚顺人，讲师，博士研究生，主要从事机械故障诊断方面的研究，E—mail：ygli@mail．neu．edu．cn
