Two Channel Quadrature Mirror Filter Design: A Sine Cosine Algorithm Based Approach

Abstract

During the last few decades multi-rate signal processing and filter banks pay very much attention to the researchers in the field of signal processing. Quadrature mirror filter banks (QMF) a kind of two channel filter banks are widely used today in the many areas of signal processing like low frequency biomedical signal processing, image processing, speech processing and communication. In order to design the two channel QMF bank, the coefficients of the filter banks are to be optimized. In this paper a new optimized method known as Sine Cosine Algorithm (SCA) is used to optimize the coefficients of prototype filter bank. The prototype filter bank is considered to get the response characteristics of ideal system in line with the response characteristics of the prototype filter bank. The SCA is used to optimize the fitness function which is formed with the help of square error at the pass band, square error at the stop band and filter bank response at the quadrature frequency. The Sine Cosine Algorithms (SCA) is a recently proposed optimization algorithm where sine and cosine are the two trigonometric functions are used. It is one global optimization which is used to modify a set of candidate solutions. The optimization through SCA, balances between exploration and exploitation to find the optimal solution very fast by minimizing the fitness function. The performance of the proposed method is analysed through mean square error (MSE) in the pass band and stop band and reconstruction error (RE). A comparison with the other optimization techniques is also presented in this paper. Two design examples are considered to prove the efficacy of the proposed algorithm.