Abstract:
Estimating the frequencies of sinusoids based on the weighted total least squares method allows for obtaining more accurate estimates compared to the ordinary least squar...View moreMetadata
Abstract:
Estimating the frequencies of sinusoids based on the weighted total least squares method allows for obtaining more accurate estimates compared to the ordinary least squares method. The most common algorithm for solving the total least squares problem is the algorithm based on the singular value decomposition (SVD) of a matrix. This algorithm has a high computational complexity. An alternative method for solving the total least squares problem is the biased normal systems approach. The paper proposes augmented system of linear algebraic equations equivalent to a weighted normal biased system of equations. and an estimate of the frequencies of sinusoids from discrete measurements with based on the proposed augmented system. The simulation results show that the frequency estimates based on the augmented systems are comparable in accuracy with the solution of the total least squares problem based on the singular value decomposition of the matrix.
Published in: 2020 IEEE East-West Design & Test Symposium (EWDTS)
Date of Conference: 04-07 September 2020
Date Added to IEEE Xplore: 15 October 2020
ISBN Information: