[西安电子科技大学阔永红现代数字信号处理英文版课件]AdvancedDigitalSignalProcessing_ch4PSE（可编辑）.docVIP

Advanced Digital Signal Processing Modern Digital Signal Processing Chapter 4 Power Spectrum Estimation  4.1 Introduction Wiener-Khintchine Theorem PSD of Ergodic Stationary Random Signal Power Spectrum Estimation PSE Estimating the PSD of real ergodic stationary random signal with finite observations sample Classical Modern PSE Methods Classical linear PSE PSE based on Fourier transform, non-parametric model method Modern nonlinear PSE PSE based on signal model parametric model method Desirable properties of PSE Unbiased Consistent Efficient High frequency resolution Narrow main lobe, low side lobe Small sample length 4.2 Classical PSE Blackman-Tukey BT Method for PSE Periodogram Method for PSE Periodogram The periodogram PSE is an asymptotically unbiased but not a consistent estimation Limitations of Classical PSE Low frequency resolution Caused by the effects of data window: 1 The degradation in resolution by main lobe; 2 The power leakage by side lobe inter-spectrum interference . Inconsistent estimation The PSE of Harmonic Process With Periodogram The PSE of White Noise With Periodogram Modifications of Classical PSE Achieving low variance at the expense of bias and frequency resolution Averaging Periodogram Bartlet method Modified Periodogram Averaging modified periodogram Welch method The PSE of Harmonic Process With Welch The sequence is divided into eight sections with 50% overlap, each section is windowed with a Hamming window The PSE of White Noise With Welch The sequence is divided into eight sections with 50% overlap, each section is windowed with a Hamming window 4.3 Parameter Model Methods for PSE Basic Principles The Time Series Model of Stationary Random Signal MA q model all-zero model Suitable for signals whose power spectra have vales but no peaks AR p model all-pole model, most widely used Suitable for s