Computationally Efficient and Noise Robust DOA and Pitch Estimation

Karimian-Azari, Sam; Jensen, Jesper Rindom; Christensen, Mads Græsbøll

doi:10.1109/taslp.2016.2577501

Cited by 10 publications

(6 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For this reason traditionally approximations with serial expansions are employed. Truncation to the first term of the serial expansion of the output signal, combined with assumption of high "signal-to-noise" (s/n) ratio, often results in acceptable predicted s/n ratios [12] - [15]. However, consideration of next expansion terms is obviously needed for essentially non-linear signals [16].…”

Section: Literature Reviewmentioning

confidence: 99%

Statistical analysis of harmonic signals for testing of Electronic Devices

Herasimov¹

2020

IJETER

View full text Add to dashboard Cite

Harmonic signals are often used for control of functional state of electronic devices. In this paper we show that the goal can be achieved via monitoring of the statistical behavior of harmonic signals propagating through the tested device. Malfunctions of the device lead to the changes in the statistical characteristics of the harmonic signals and thus are used for the monitoring of the technical state of the device. We consider statistical behavior of such signals at the variation of their phase, caused by malfunctions in electronic devices. Deviation of the mathematical expectation from its nominal values and dispersion are proposed for estimation of malfunction of a device. A functional scheme of a controller analyzing the mathematical expectation and dispersion is proposed. An experiment demonstrating how the proposed approach can be used for monitoring of the technical state of an electronic device under the influence of the external noise is presented.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Statistical analysis of harmonic signals for testing of Electronic Devices

Herasimov¹

2020

IJETER

View full text Add to dashboard Cite

show abstract

“…The problem of estimating these parameters from noisy observed signals is considered beyond the scope of this paper, and we instead refer the interested readers to the many existing methods for finding them, e.g. [24,25,26,27,28,23]. We design fixed delay-and-sum and null forming beamformers herein with distortionless constraints based on the aforementioned spatial and spectral parameters of the multichannel signals.…”

Section: Introductionmentioning

confidence: 99%

Harmonic beamformers for speech enhancement and dereverberation in the time domain

Jensen

Karimian-Azari

Christensen

et al. 2020

Speech Communication

View full text Add to dashboard Cite

“…Therefore, according to the Shannon-Kotelnikov theorem, at any finite sampling rate, accurate reconstruction of the signal is impossible, because the high-frequency components of the output signal cannot be restored. In articles [7,8] the known traditional methods of recovery of an analog signal from discrete have resulted. These methods consist of the sequential solution of two problems: recovery of the output signal of the ) t ( f EAU from the discrete signal q f at the output of the EDU and recovery of the input signal of the ) t ( g EAU from its output signal ) t ( f .…”

Section: Literature Analysis and Problem Statementmentioning

confidence: 99%

“…Let us express the eigenvalues ) (   in terms of the frequency response ) , ( H   . To do this, we use relation (7), replacing it x with ) x ( − and dividing the integration interval into sections t  :…”

Section: The Proposed Mechanismmentioning

confidence: 99%

The method of discretization signals to minimize the fallibility of information recovery

Лаптєв

Yevseiev

Hatsenko

et al. 2021

Int. j. commun. netw. inf. secur.

View full text Add to dashboard Cite

The paper proposes a fundamentally new approach to the formulation of the problem of optimizing the discretization interval (frequency). The well-known traditional methods of restoring an analog signal from its discrete implementations consist of sequentially solving two problems: restoring the output signal from a discrete signal at the output of a digital block and restoring the input signal of an analog block from its output signal. However, this approach leads to methodical fallibility caused by interpolation when solving the first problem and by regularizing the equation when solving the second problem. The aim of the work is to develop a method for the signal discretization to minimize the fallibility of information recovery to determine the optimal discretization frequency.The proposed method for determining the optimal discretization rate makes it possible to exclude both components of the methodological fallibility in recovering information about the input signal. This was achieved due to the fact that to solve the reconstruction problem, instead of the known equation, a relation is used that connects the input signal of the analog block with the output discrete signal of the digital block.The proposed relation is devoid of instabilities inherent in the well-known equation. Therefore, when solving it, neither interpolation nor regularization is required, which means that there are no components of the methodological fallibility caused by the indicated operations. In addition, the proposed ratio provides a joint consideration of the properties of the interference in the output signal of the digital block and the frequency properties of the transforming operator, which allows minimizing the fallibility in restoring the input signal of the analog block and determining the optimal discretization frequency.A widespread contradiction in the field of signal information recovery from its discrete values has been investigated. A decrease in the discretization frequency below the optimal one leads to an increase in the approximation fallibility and the loss of some information about the input signal of the analog-to-digital signal processing device. At the same time, unjustified overestimation of the discretization rate, complicating the technical implementation of the device, is not useful, since not only does it not increase the information about the input signal, but, if necessary, its restoration leads to its decrease due to the increase in the effect of noise in the output signal on the recovery accuracy. input signal. The proposed method for signal discretization based on the minimum information recovery fallibility to determine the optimal discretization rate allows us to solve this contradiction.

show abstract

Computationally Efficient and Noise Robust DOA and Pitch Estimation

Cited by 10 publications

References 44 publications

Statistical analysis of harmonic signals for testing of Electronic Devices

Statistical analysis of harmonic signals for testing of Electronic Devices

Harmonic beamformers for speech enhancement and dereverberation in the time domain

The method of discretization signals to minimize the fallibility of information recovery

Contact Info

Product

Resources

About