Transform domain LMS algorithms for sparse system identification

Shi, Kun; Ma, Xiaoli

doi:10.1109/icassp.2010.5495882

Cited by 19 publications

(11 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The most frequently used adaptive algorithms are the ones based on the least mean square (LMS) algorithm (Feintuch, 1976;Widrow and Stearns, 1985). Several improvements have been reported to either enhance the convergence or to improve the accuracy of modeling (Lee and Gan, 2004;Gu et al, 2009;Chen et al, 2009a;Shi and Ma, 2010;Yu et al, 2014b).…”

Section: Introductionmentioning

confidence: 96%

Swarm and evolutionary computing algorithms for system identification and filter design: A comprehensive review

Gotmare

Bhattacharjee

Patidar

et al. 2017

Swarm and Evolutionary Computation

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 96%

Swarm and evolutionary computing algorithms for system identification and filter design: A comprehensive review

Gotmare

Bhattacharjee

Patidar

et al. 2017

Swarm and Evolutionary Computation

View full text Add to dashboard Cite

“…The proposed ZA-MP-CFxLMS algorithm can be generalized for other sparse CFxLMS algorithms, and the proposed ZA-WD-CFxLMS algorithm can be generalized for other transfer domain (TD) sparse CFxLMS algorithms. These are novel additions to the existing theory as current sparsity solutions are limited to their non-complex variants [24,25].…”

Section: Introductionmentioning

confidence: 99%

Sparse complex FxLMS for active noise cancellation over spatial regions

Zhangg

Abhayapala

Samarasinghe

et al. 2016

2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

In this paper, we investigate active noise control over large 2D spatial regions when the noise source is sparsely distributed. The 1 relaxation technique originated from compressive sensing is adopted and based on that we develop the algorithm for two cases: multipoint noise cancellation and wave domain noise cancellation. This results in two new variants (i) zero-attracting multi-point complex FxLMS and (ii) zero-attracting wave domain complex FxLMS. Both approaches use a feedback control system, where a microphone array is distributed over the boundary of the control region to measure the residual noise signals and a loudspeaker array is placed outside the microphone array to generate the anti-noise signals. Simulation results demonstrate the performance and advantages of the proposed methods in terms of convergence rate and spatial noise reduction levels.

show abstract

“…Several sparsity-aware modifications of the standard LMS have been introduced in the literature [7,8,9,10,11,12,13,24].…”

Section: Standard Lmsmentioning

confidence: 99%

“…A performance analysis of the l 0 -pseudo-norm constraint LMS algorithm of [8] is given in [10]. In [11], [12], variations of the algorithms in [7] are introduced. In [11], the filter coefficients are updated in a transform domain which leads to faster convergence for non white inputs.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Reweighted l1-norm penalized LMS for sparse channel estimation and its analysis

Taheri

Vorobyov

2014

Signal Processing

View full text Add to dashboard Cite

A new reweighted l 1 -norm penalized least mean square (LMS) algorithm for sparse channel estimation is proposed and studied in this paper. Since standard LMS algorithm does not take into account the sparsity information about the channel impulse response (CIR), sparsity-aware modifications of the LMS algorithm aim at outperforming the standard LMS by introducing a penalty term to the standard LMS cost function which forces the solution to be sparse. Our reweighted l 1 -norm penalized LMS algorithm introduces in addition a reweighting of the CIR coefficient estimates to promote a sparse solution even more and approximate l 0 -pseudo-norm closer. We provide in depth quantitative analysis of the reweighted l 1 -norm penalized LMS algorithm. An expression for the excess mean square error (MSE) of the algorithm is also derived which suggests that under the right conditions, the reweighted l 1 -norm penalized LMS algorithm outperforms the standard LMS, which is expected. However, our quantitative analysis also answers the question of what is the maximum sparsity level in the channel for which the reweighted l 1 -norm penalized LMS algorithm is better than the standard LMS. Simulation results showing the better performance of the reweighted l 1 -norm penalized LMS algorithm compared to other existing LMS-type algorithms are given.

show abstract

Transform domain LMS algorithms for sparse system identification

Cited by 19 publications

References 6 publications

Swarm and evolutionary computing algorithms for system identification and filter design: A comprehensive review

Swarm and evolutionary computing algorithms for system identification and filter design: A comprehensive review

Sparse complex FxLMS for active noise cancellation over spatial regions

Reweighted l1-norm penalized LMS for sparse channel estimation and its analysis

Contact Info

Product

Resources

About