论文部分内容阅读
本文提出了一种人工神经网络自适应ＩＩＲ滤波器，这种自适应ＩＩＲ滤波器采用并联型结构，用人工神经网络实现，并保证系统在自适应过程中的稳定性，从而得到了一种稳定的、高度并行的自适应ＩＩＲ滤波器，从根本上改变了以往的串行数值迭代系统，使滤波器自适应过程仅需要几个微秒就可以完成。从而有可能用神经自适应系统完成对快速变化信号的实时处理。本文给出了计算机模拟的结果，理论和模拟结果均表明该结构是稳定的，收敛速度也有明显增加。
In this paper, an artificial neural network adaptive IIR filter is proposed. The adaptive IIR filter adopts a parallel structure, which is implemented by artificial neural network and ensures the stability of the system in the process of self-adaptation. Thus, a stable The highly parallel adaptive IIR filter has fundamentally changed the previous serial numerical iteration system so that the filter adaptation process takes only a few microseconds to complete. It is therefore possible to implement real-time processing of rapidly changing signals using a neural adaptive system. In this paper, the results of computer simulation are given. The theoretical and simulation results show that the structure is stable and the convergence rate also increases obviously.