一种新的FIR滤波器脉动实现结构

A Novel Structure of Systolic Implementation for FIR Filter

为了提高ＦＩＲ滤波器的处理速度，一个主要手段是并行处理技术．并行处理除了可以提高运算速度外，还可以提高ＦＩＲ滤波器的数据通过率以及降低系统功耗．本文首先从多项式分解角度给出一种ＦＩＲ滤波器的并行结构．通过对此并行ＦＩＲ滤波器的分析，提出了一种新的ＦＩＲ滤波器的脉动实现结构．这种结构与一般的实现ＦＩＲ滤波器的脉动结构相比具有规模小、能适应更高处理速度的优点．

Parallel processing technique is a main method to be used to increase the processing speed of the FIR filters. It can also be used to improve the data throughput and reduce the consumed power by the system.At first,a parallel structure of the FIR filter is presented based on polynomial decomposition method. A novel structure of systolic implementation for the FIR filter is presented by analyzing the parallel FIR filter structure. A comparison of this novel structure with the ordinary systolic implementation of the FIR filter shows that this structure has a small scale and can be suitable to the cases which need higher processing speed.