Lecture Notes on the Relationship between the Power Spectrum Estimated by MVDR and CBF
Qun Wan, Jingcheng Qi, Qiang Yan, Lin Zou, Xian-Sheng Guo · 2019
Power spectrum density estimation is a quite important tool in signal processing, from which we can analyze the frequency components of the random signal. CBF and MVDR are classical and modern power spectrum estimation methods respectively. Both of them can achieve ideal performance under certain conditions. Here we discuss the relationship between MVDR and CBF on the power spectrum function. Interestingly, we found that when the observation vector contains no more than two frequency components, CBF and MVDR power spectrum will peak at the same frequencies, and the peaks are the same. We hope that the results of this lecture notes can give some inspiration to students who are studying digital signal processing.