基于多层感知器的3型多带通FIR滤波器设计

Design of Multi-Band-Pass FIR Filters of Type-3 Based on MLP Network

利用多层感知器神经网络设计了3型多带通FIR滤波器,其基本思想是该神经网络利用总误差的能量函数,使能量函数递减,然后根据给定的幅频响应以及通带和阻带误差设计滤波器。提供了有效的训练方法,克服了传统方法的主要缺陷,改进了滤波器的设计。该方法设计滤波器收敛速度很快,适用于高阶系统。仿真结果表明该方法设计的3型FIR线性相位多带通滤波器的性能接近于理想滤波器。

A multilayer perceptron network for the design of the FIR multi-band-pass filters of type-3 are proposed. The network is contrived to have an energy function that coincides with the sum-squared error of the approximation problem at hand and by ensuring that the energy is a monotonic decreasing function, the approximation problem can be solved. Two solutions are obtained. In the first the FIR band-pass filter of type-3 is designed on the basis of a specified amplitude response and in the second a filter that has specified maximum passband and stopband errors is (designed.) The training method is proposed, and design of the filters is improved. It overcomes the main disadvantages of the conventional methods. Furthermore, the convergence speed is fast, and it is also suitable to the design of high order systems. The simulation results show that the FIR multi-band-pass filters of type-3 designed using the MLP network have the performances of ideal filters.