Application Of Ant Colony Algorithm In The Optimal Design Of FIR Digital Filters

As a basic function unit of digital signal processing, finite impulse response (FIR) digital filters have been widely used in the fields of electronic information such as digital signal processing. If a floating-point arithmetic is used in the digital signal processor, the wordlength effect can nearly always be neglected. However, for a fast implementation of a digital filter, a fixed-point arithmetic has to be used and the wordlength effect must be considered. In order to improve the filters'performance that is reduced by the wordlength effect, an algorithm should be applied to optimally design the filter coefficients. The optimal design of FIR filters with linear phase based on ant colony algorithm (ACA) is discussed in this dissertation, which is organized as follows:(1) Firstly, the background and significance of FIR digital filters optimal design are introduced; also the survey of interrelated domain aboard and interiorly is mentioned. Secondly, optimality criteria and optimal design methods of FIR digital filters are briefly discussed. The basic principle of ant colony algorithm (ACA) and its mathematical models and application are summarized.(2) Based on the full analysis of the basic ant colony algorithm, ACA is applied to determine the fixed-point coefficients in the optimal design of FIR filters with linear phase. Furthermore, ACA is correspondingly modified according to the characteristic of the problem on FIR digital filters. For a FIR digital filter, two different coefficients may be equal to each other; therefore the Tabu list is cancelled. The number of ants is set to equal the number of coefficients to be considered and ants are destined to cooperatively search better values for the coefficients, which can avoid the method being trapped because of the cancellation of Tabu list. The procedures for designing FIR digital filters by ACA are given and different optimality criteria are considered. We compared the performance of the optimally designed FIR digital filters with other existing algorithms, and the feasibility and effectiveness of ACA are certified by the simulation results.(3) The parameter selection is a key factor of affecting the performance and efficiency of the ACA. On the basis of brief introduction of model for FIR digital filters optimal design based on ACA, we analyzes the influence of the parameters involved in the performance of the proposed method, derives the basic formula of parameters configuration, and presents the instructional principle of parameter selection. Simulation demonstrates that the parameters configuration principle is effective for improving optimal design performance of FIR digital filters and beneficial to the application of the ACA in FIR digital filters optimal design.It is of certain significance in both theory and engineering for this work to exploit a new approach for the design and implementation of FIR digital filter with linear phase.