基于QPSO的两通道正交镜像滤波器组的优化设计*

通过分析两通道正交镜像滤波器组的重构条件,提出了一种新的两通道正交镜像滤波器(QMF)组的设计方法,将完全重构的两通道QMF组的设计问题转换为由原型低通滤波器在通带、阻带和过渡带幅频响应的均方误差加权和组成的一种无约束的非线性优化问题,用量子粒子群优化算法(QPSO)来有效求解该类优化问题,获得该问题的全局最优解.最后...

Two-channel QMF bank optimization design based on quantum-behaved particle swarm optimization algorithm

This paper analyzed the reconstruction condition of two-channel QMF bank, proposed an new method to design two-channel QMF bank. At first, the design problem of perfect reconstruction two-channel QMF had been recasted into a unconstrained nonlinear optimization problem, which was made up of the weighted sum of the mean square error of amplitude response of prototype low-pass filter in passband, stopband and transition. Secondly, proposed QPSO to efficiently tackle the kind of optimization problem, then got the global optimal solution. Finally, designed a two-channel QMF bank by the obtained optimal coefficients. Educed the conclusion through experiments and comparisions, that the new method is simpler and more effective than conventional method under satisfying performance, the reconstruction error of QMF banks is lower .