论文部分内容阅读
提出一种基于改进差分进化算法和开关电流微分器的连续小波变换实现的新方法。引入混沌初始化种群、自适应变异和指数递增交叉概率因子,提高差分进化算法的收敛速度和精度,构造小波函数逼近模型,并利用改进差分进化算法,求得小波逼近函数及频域传递函数。采用开关电流微分器作为基本单元,设计冲激响应为小波逼近函数的开关电流滤波器组,通过调节小波滤波器的时间常数,获得多尺度小波函数,实现连续小波变换。实验结果表明该方法可行。
A new method of continuous wavelet transform based on improved differential evolution algorithm and switched current differentiator is proposed. The chaotic initialization population, adaptive mutation and exponential increasing crossover probability factor are introduced to improve the convergence speed and accuracy of the differential evolution algorithm. The wavelet function approximation model is constructed and the wavelet approximation function and frequency transfer function are obtained by the improved differential evolution algorithm. Using switch current differentiator as the basic unit, a switching current filter bank with impulse response as wavelet approximation function is designed. By adjusting the time constant of wavelet filter, multi-scale wavelet function is obtained and continuous wavelet transform is realized. Experimental results show that this method is feasible.