A new quasi-Newton adaptive filtering algorithm

Campos, Marcello L. R. de; Antoniou, A.

doi:10.1109/82.644046

Cited by 44 publications

(52 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The QN algorithm proposed in [4] is a robust algorithm that was developed using a stochastic approach and is based on the rank-one quasi-Newton update of the Hessian matrix, i.e., satisfying the QN hereditary condition [7]. However, this algorithm can also be shown to minimize the following deterministic cost function:…”

Section: Qn Algorithm and The Minimum-disturbance Approachmentioning

confidence: 99%

“…Robust RLS algorithm implementations with reduced computational complexity are usually based on QR decompositions, which are complex to implement and maintain [1]. There are other algorithms that have been developed based on known convex optimization methods, like the quasi-Newton (QN) [4] and interior point least squares (IPLS) algorithms [6]. These algorithms offer increased robustness at a cost of extra computational complexity, for they do not admit O(N ) implementations.…”

Section: Introductionmentioning

confidence: 99%

“…Lack of robustness may be due to accumulation of quantization errors or due to loss of positive definiteness of the Hessian matrix caused by nonpersistently exciting input signals [4], [5].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Minimum-disturbance description for the development of adaptation algorithms and a new leakage least squares algorithm

Castoldi

Campos

2009

2009 IEEE International Conference on Acoustics, Speech and Signal Processing

View full text Add to dashboard Cite

Usual methods for the development of adaptive filters are based on a stochastic approximation of the gradient vector and Hessian matrix, or on a deterministic minimization of quadratic a posteriori output errors. Gradient-based algorithms are usually placed in the first group, whereas least squares (LS) based algorithms are placed in the second group. These are just how algorithms are usually presented and analyzed and alternative descriptions exit. This paper proposes to shed new light onto known adaptation algorithms by means of a minimum-disturbance approach to the cost function together with constraints added to improve their robustness. The resulting algorithms are able to perform extremely well in many demanding applications.

show abstract

Section: Qn Algorithm and The Minimum-disturbance Approachmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Minimum-disturbance description for the development of adaptation algorithms and a new leakage least squares algorithm

Castoldi

Campos

2009

2009 IEEE International Conference on Acoustics, Speech and Signal Processing

View full text Add to dashboard Cite

show abstract

“…The model was used for analyzing the NLMS algorithm [12], and was shown to yield accurate results. The model was also successfully used to analyze the quasi-Newton (QN) [14] and the binormalized data-reusing LMS (BNDRLMS) [15] algorithms.…”

Section: Convergence Analysismentioning

confidence: 99%

“…For the MSE analysis, we assume that the vectors are excited in a discrete number of directions [12,14,15].…”

Section: Excess Mse For White Input Signalsmentioning

confidence: 99%

Mean-Squared Analysis of the Partial-Update NLMS Algorithm

Werner¹,

Campos²,

Diniz³

2003

Journal of Communication and Information Systems

View full text Add to dashboard Cite

-In this paper, we present mean-squared convergence analysis for the partial-update normalized least-mean square (PU-NLMS) algorithm with closed-form expressions for the case of white input signals. The analysis uses order statistics and the formulas presented here are more accurate than the ones found in the literature for the PU-NLMS algorithm. Simulation results show excellent agreement with the the results predicted by the analysis.Keywords: Adaption algorithms, efficient algorithms, partial update, MSE analysis, order statistics, NLMS algorithm.Resumo -Neste artigo apresentamos a análise da convergência na média quadrática do algoritmo least-meansquare normalizado com utilização parcial dos dados (PU-NLMS) para o caso de sinais de entrada brancos. A análise usa conceitos de estatística de variáveis ordenadas e as fórmulas apresentadas mostraram-se mais precisas que outros resultados encontrados na literatura para o algoritmo PU-NLMS. Resultados de simulações mostramótima concordância com aqueles previstos pela teoria. Palavras-chave:Algoritmos adaptativos, algoritmos efficientes, algoritmos com atualização parcial de dados, estatítica de variáveis ordenadas, algoritmo LMS normalizado.

show abstract

Set‐theoretic adaptive filtering based on data‐driven sparsification

Yukawa¹,

Yamada²

2011

Adaptive Control & Signal

View full text Add to dashboard Cite

In this article, we propose a fast and efficient algorithm named the adaptive parallel Krylov-metric projection algorithm. The proposed algorithm is derived from the variable-metric adaptive projected subgradient method, which has recently been presented as a unified analytic tool for various adaptive filtering algorithms. The proposed algorithm features parallel projection-in a variable-metric sense-onto multiple closed convex sets containing the optimal filter with high probability. The metric is designed based on (i) sparsification by means of a certain data-dependent Krylov subspace and (ii) maximal use of the obtained sparse structure for fast convergence. The numerical examples show the advantages of the proposed algorithm over the existing ones in stationary/nonstationary environments.

show abstract

A new quasi-Newton adaptive filtering algorithm

Cited by 44 publications

References 16 publications

Minimum-disturbance description for the development of adaptation algorithms and a new leakage least squares algorithm

Minimum-disturbance description for the development of adaptation algorithms and a new leakage least squares algorithm

Mean-Squared Analysis of the Partial-Update NLMS Algorithm

Set‐theoretic adaptive filtering based on data‐driven sparsification

Contact Info

Product

Resources

About