基于协调粒子群算法的无限冲激响应滤波器优化设计  

An Effective Optimization Method for Designing IIR Digital Filters Based on AFHPSO

作　　者：甘明刚[1] 潘峰[1] 蔡涛[1] 

机构地区：[1]北京理工大学自动化学院,北京100081

出　　处：《兵工学报》2010年第10期1398-1402,共5页Acta Armamentarii

基　　金：北京市教育委员会共建基金项目(1000702)

摘　　要：确定滤波器的参数和阶数是无限冲激响应(IIR)数字滤波器设计中的最重要问题。目前的IIR滤波器优化设计方法容易陷入局部极值,难以得到全局最优解,且阶数不易确定。针对IIR数字滤波器优化设计实质上是一个对参数优化的问题,引入了协调粒子群算法对IIR滤波器进行优化设计,将传统设计方法和优化设计方法相结合,根据传统设计方法确定滤波器的阶数,通过协调粒子群算法对参数进行优化,将最小均方误差函数作为适应度函数,将滤波器的性能要求作为算法终止条件,滤波器的稳定性作为约束条件。该方法解决了传统方法设计的结果不理想以及常用优化设计方法不收敛且阶数不易确定的问题。经过在实际系统中的应用,表明该方法是有效、实用的。The optimization design of IIR digital filters is a multi-parameter problem.A satisfactory optimization method based on acceleration factor harmonious PSO（AFHPSO） was proposed to design IIR digital filters.The method was combined by traditional and optimization methods.The rank of IIR filter was computed by traditional way and the parameters were optimized.The minimum mean-square error was used as the fitness function,the performance requirement was used as the termination condition,and the stabilization of filter was used as a constrained condition in the method.The disadvantage of the other methods was improved in the method,for example,the rank is hard to computed in the generic optimization method.The result shows this is an effective and practical method.

关 键 词：控制理论 无限冲激响应数字滤波器 设计 优化 协调粒子群算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TN911.7[自动化与计算机技术—控制科学与工程]