神经网络在高阶滤波器优化设计中的应用研究

Application of Neural Networks in the Optimized Design of High-order Filters

本文详细研究了FIR线性相位滤波器的幅频特性与余弦基函数神经网络算法的关系，证明了神经网络系统的稳定条件，给出了FIR高阶滤波器优化设计实例。计算机仿真结果表明了该算法在高阶滤波器设计中的有效性和优异性能。

Relationship between the amplitude-frequency characteristic of the FIR linear phase filters and the algorithm of neural networks on cosine basis functions is extensively studied. Stability conditions for the neural network systems are presented and proven. Examples of the optimized design on the High-order FIR filters are also given. Experiment results show that the algorithm is efficient and effective for the design of the high-order filters.