自适应格型IIR滤波器的进化学习算法

Lattice Form Adaptive IIR Filtering with Evolutionary Learning Algorithm

本文提出了自适应全零点格型ＩＩＲ滤波器结构及进化学习算法。全零点格型结构的自适应ＩＩＲ滤波器（ＡＩＩＲＦ）具有结构简单，计算量少等优点．新的滤波器结构和进化学习算法相结合，不仅解决了ＡＩＩＲＦ的稳定性问题，而且滤波器参数能够以很大的概率快速收敛于全局最优解，同时允许大动态范围的输入信号，尤其对高阶、极点靠近单位圆的ＡＩＩＲＦ。计算机仿真结果验证了该滤波器的有效性．

This paper presents a new lattice adaaptive IIR filter (AIIRF) using evolutionary learning algorithm. The new lattice AIIRF is with simple structure and less compution burden. Evoluttionary learning algorithm for AIIRF problme provides graranteed fiber stability. AIIRF using evolutionary learning algorithm can almost sure reach the global optimization of the objective function of AIIRF and quickly converge. Numerical simulations show that the new AIIRF has advantage in the case where poles of filter spstem are close to the unit circle and for high-order the problems.