Adjustable window based design of multiplier-less cosine modulated filter bank using swarm optimization algorithms

In this paper, an adjustable window based approach is presented for the design of multiplier-less near perfect reconstruction cosine modulated filter bank (CMFB) for specified stopband attenuation (As) and channel overlapping. Kaiser window function is employed for designing the computationally efficient prototype filter with filter coefficients in canonic signed digit (CSD) space. Optimized performance of the designed filter is achieved using swarm based algorithm such as cuckoo search (CS) optimization, so that the filter coefficients of a multiplier-less prototype filter are optimized to achieve the magnitude response of 0.707 at frequency ω = π/2M. In this method, instead of using two optimization techniques: one for designing continuous coefficients and other for optimizing quantized prototype filter coefficients, single swarm based technique is used, while a comparative study using proposed scheme based on the performance of different window functions as well as different swarm based techniques such particle swarm optimization (PSO) and artificial bee colony (ABC) algorithm is made. Design examples presented, using this technique, illustrates the improved performance of proposed technique as compared to other published algorithms.