裁剪频谱的线性系统DFT反卷积算法

在对传统求冲击响应方法的误差进行分析的基础上，提出了一种新的卷积反演算法。该算法通过提高采样率后对频谱进行裁剪，然后运用zoom运算降低采样率，再通过反卷积求出系统的冲击响应。仿真实验证明，该算法能有效地避免量化误差的放大和迁移所造成的影响，比传统方法和最小二乘法具有更小的误差。

Linear System DFT Deconvolution Algorithm Based on Frequency Spectrum Tailoring

Based on the error analysis of the traditional algorithm used to solve the impulse response in a linear system, a new deconvolution algorithm is proposed. With the increased sampling frequencies, the spectrum is tailored and zoom operation is used to decrease the sampling frequencies. Finally, the impulse response is obtained by deconvolution. Simulations demonstrate that the algorithm can reduce the quantification sampling error and the errors amplified and delivered in the procedure of computation. The method has a better precision than the traditional algorithm and the least square algorithm.