论文部分内容阅读
针对Goertzel算法的计算结构硬件实现效率低等问题,提出了利用两个FIR滤波器实现Goerztel算法,并且将改进后的计算结构由原来的AR(2)过程推广到一般的AR(p)过程.改进后的计算结构避免了递推运算,能够预确定数据的动态范围进行定标,充分利用了DSP等信号处理器的硬件结构,适合定点处理器的编程.通过在定点DSP处理器TMS320C5510上进行软件仿真,测试结果表明改进计算结构后算法的效率是改进前算法效率的3.7倍.
In order to solve the problem of low efficiency of Goertzel algorithm in computing hardware architecture, this paper proposes a Goerztel algorithm using two FIR filters and extends the original AR (2) process to the general AR (p) process. The improved calculation structure avoids the recursive computation and can pre-determine the dynamic range of the data to be calibrated, and makes full use of the hardware structure of the signal processor such as DSP and is suitable for the programming of the fixed-point processor.Through the fixed-point DSP processor TMS320C5510 Simulation results show that the efficiency of the algorithm is improved by 3.7 times the efficiency of the former algorithm.