A canonic-signed-digit coded genetic algorithm for designing finite impulse response digital filter

作者：Pan Shing Tai^*

来源：Digital Signal Processing, 2010, 20(2): 314-327.

DOI：10.1016/j.dsp.2009.06.024

• Canonic signed digit (CSD)
• Genetic algorithm
• Finite-impulse filter (FIR)

In this paper, the genetic algorithm (CA) based on Canonic Signed Digit (CSD) code was used to find the optimum design of a finite impulse response digital filter (FIR). By using the characteristics of the CSD structure. the circuit was able to be simplified and also the calculation speed was raised to increase the hardware's efficiency. However, CSD structure cannot be guaranteed by a general CA after the evolution of chromosomes. Thus in this research an algorithm was proposed which the CSD structure can be maintained. A CSD coded CA was used to the evolution of chromosome to reduce the time wasted by trials and errors during the evolution and then to accelerate the training speed. In this paper, a new hybrid code for the filter coefficients was proposed to improve the precision of the coefficient of FIR. Ail example is shown in this paper to verify the efficiency of the proposed algorithm.

• 出版日期2010-3