Fast and Efficient Estimation of Frequencies

Parameter estimation of multiple sinusoids is one of the fundamental problems of signal processing having numerous application areas, including wireless communications and array signal processing. This article proposes an iterative, fast, and efficient algorithm to estimate the frequencies and ampli...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on communications 2021-06, Vol.69 (6), p.4054-4066
Main Author: Serbes, Ahmet
Format: Article
Language:English
Subjects:
Algorithms
Computational efficiency
direction of arrival estimation
Discrete Fourier transforms
Doppler measurements
Estimation
Frequency estimation
Iterative methods
Lower bounds
Parameter estimation
Signal processing
Signal processing algorithms
Signal to noise ratio
Sine waves
spectral analysis
Wireless communications
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Parameter estimation of multiple sinusoids is one of the fundamental problems of signal processing having numerous application areas, including wireless communications and array signal processing. This article proposes an iterative, fast, and efficient algorithm to estimate the frequencies and amplitudes of multiple sinusoids in noisy environments. The proposed approach interpolates on the q -shifted DFT coefficients of multiple sinusoids. We theoretically demonstrate that the performance of the proposed algorithm depends on DFT-shift, number of iterations and minimum DFT frequency separation parameters. Further, mean square error of the proposed algorithm achieves the Cramér-Rao lower bound when these parameters are selected appropriately. We also propose solid bounds regarding how these parameters can be selected for optimum overall performance. The total computational cost of the proposed algorithm is in the order of \mathcal {O}(K N\log N) , where K is the number of sinusoidal components and N is the signal length. We provide comprehensive numerical simulations that confirm the correctness of our theoretical deductions.
ISSN:0090-6778
1558-0857
DOI:10.1109/TCOMM.2021.3067383