Hybrid particle swarm optimization design of cosine modulated filter banks

Modulated filter performance mainly depends on characteristics of prototype filter. The key to modulated filter banks is to design high-performance prototype filter. A new design method of cosine modulated filter banks is presented, which uses polynomial approximation method and particle swarm optimization method to design prototype low-pass filter and then gets analysis and synthesis filter after cosine modulation. Finally, the proposed method is illustrated by a simulation example by using MATLAB. The experimental results indicate that filter banks designed by the given method have good stop-band attenuation characteristics and the given method has high practical value in filter banks designing.