算术傅立叶变换中非均匀采样点信号值的一种分段估计法

Space-ividing Method to Estimate Signal Values at Non equidistant Sampling Points

提出了一种用于估计算术傅立叶变换 ( AFT)中非均匀采样点处信号值的方法 .这种方法在不影响原算法高速性及适于 VLSI实现这一特性的基础上 ,改善了 AFT算法在零阶插值时的误差 .基本原理是在两均匀采样点之间根据某种原则进行分段 ,每个分段区间赋予一个只需简单计算的值 .仿真表明 :在较少分段时 ,分段 AFT即可逼近一阶插值 AFT的误差 ,同时其误差效果也好于迭代实现的自适应

A segmentation method to estimate signal values at non equidistant sampling points was presented. Its aim is the error improvement in the 0 order arithmetic Fourier transform algorithm with a small increase on computations that even can be negligible. The simulations show that the errors in these estimation methods are very close to the errors in the 1 order interpolation estimation method that has much multiplication.