进化论动态参量自适应消噪算法

由于以输出信号平均能量的倒数为适应度函数、以及定值克隆系数和定值匹配系数的因素,制约了传统进化论自适应消噪算法的收敛特性,影响了传统进化论自适应算法降噪的有效性。针对该问题,提出了基于动态适应度函数的进化论变系数自适应消噪算法。新算法中,匹配系数与进化代数相关联,克隆系数由进化代数和适应度值决定,并采用动态适应度函数,改善了滤波器的收敛特性和噪声抑制能力。模拟仿真分析表明,新算法较同类进化论算法有较快的收敛速度和良好的消噪效果。物理台架实验数据验证了该算法对故障信号的有效提取能力。

A Variable Evolutionary Coefficients Adaptive Noise Cancellation Algorithm

Factors of fitness function using the reciprocal of the average energy of the output signal, constant cloning and mating coefficients all restrict the convergence and noise cancellation performances of the traditional evolutionary adaptive noise cancellation algorithm. In this paper, the variable coefficients evolutionary adaptive noise cancellation algorithm using dynamic fitness function is proposed. The mating coefficient is related to evolutionary generation and the cloning coefficient is determined by evolutionary generation and fitness value, and a dynamic fitness function is also employed. Results from simulation case and test datasets show that the proposed algorithm exhibits better convergence and noise cancellation performances than other similar algorithms.