Steady-State Analysis of the Deficient Length Incremental LMS Adaptive Networks

Azarnia, Ghanbar; Tinati, Mohammad Ali

doi:10.1007/s00034-015-9978-7

Cited by 13 publications

(13 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…, j ∈ N i being the saddle point of (20). By following the above arguments in the converse way, it results that the saddle point of (20) is the fixed point of the iterations (21)- (24).□…”

Section: General Frameworkmentioning

confidence: 92%

“…[34]): Let F: ℂ N → ( − ∞, ∞] be a convex function. Then x = P G (z) if and only if z ∈ x + ∂F(x), where ∂F is the sub-differential of F. Proof of Theorem 1: By applying Lemma 1 to (21)- (24) we have…”

Section: General Frameworkmentioning

confidence: 97%

“…So, distributed processing is often much desirable for such situations. In these scenarios, each node communicates only with its closest neighbours and processing task is carried out locally at every node, without any FC unit [24,25]. Jointly distributed sparse recovery methods are suitable for WSNs.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Cooperative and distributed algorithm for compressed sensing recovery in WSNs

Azarnia

Tinati

Rezaii

2018

IET signal process.

Self Cite

View full text Add to dashboard Cite

Wireless sensor networks (WSNs) could benefit a lot from compressive sensing (CS). Inherent physical structure of sensors of WSNs (battery-powered devices) demands computational-efficient algorithms with no heavy burden on a small subset of the sensors, i.e. fusion sensors. This could be achieved by distributed algorithms in which computation is distributed among all sensor nodes. On this basis, in this study, the authors have proposed a distributed and cooperative sparse recovery algorithm in which each sensor decodes a sparse signal by running a recovery algorithm with the cooperation of its neighbours. The proposed algorithm has a general structure and can be adapted to many optimisation algorithms in the context of the CS. This algorithm is completely distributed and requires an acceptable computational complexity that is suitable for WSNs. A detailed proof of convergence behaviour of the proposed algorithm is also presented. The superiority of the proposed algorithm compared with similar methods in terms of recovery quality and convergence rate is confirmed through simulation.

show abstract

Section: General Frameworkmentioning

confidence: 92%

Section: General Frameworkmentioning

confidence: 97%

See 1 more Smart Citation

Cooperative and distributed algorithm for compressed sensing recovery in WSNs

Azarnia

Tinati

Rezaii

2018

IET signal process.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Consequently, the previous theoretical results on the sufficient length NSAF algorithm do not necessarily apply to the deficient length situation. For such scenarios, the performance of many algorithms has been studied in the literature such as the LMS [19], [20], the frequency-domain block LMS (FBLMS) [21], and the distributed LMS [22], [23]. To the best of our knowledge, however, there are no available studies to accurately evaluate the performance of the deficient length NSAF algorithm.…”

Section: Introductionmentioning

confidence: 99%

Performance analysis of the deficient length NSAF algorithm and a variable step size method for improving its performance

Zhao

2017

Digital Signal Processing

View full text Add to dashboard Cite

A  bstract: In all presented analyses of the normalized subband adaptive filter (NSAF) algorithm, there is a common assumption that the length of the adaptive filter is equal to that of the unknown system. In many practices, however, the adaptive filter usually works in an under-modeling situation. Namely, the length of the adaptive filter is less than that of the unknown system. Therefore, for this case, the existing analysis results for the NSAF algorithm are not applicable. In this paper, we analyze the performance of the deficient length NSAF algorithm based on some reasonable assumptions and approximations. More precisely, the expressions that characterize the transient-state and steady-state mean-square behavior of the algorithm are presented. Simulation results in various scenarios support our theoretical expressions. In addition, based on our analyses, a variable step size NSAF algorithm suitable for the under-modeling case is developed, which improves the performance.

show abstract

“…The performance of the incremental LMS algorithm in this realistic case is studied. 33 The incremental combination of RLS and LMS adaptive filters 34 is explored as design solutions to enhance the overall performance of an adaptive system. By comparing the incremental LMS and the steepest-descent algorithms for distributed estimation in the case of diminishing step size, we find an interesting fact that the incremental LMS outperforms the steepest descent at the initial stage of algorithm (before convergence), while the situation is conversed in the steady state (after convergence).…”

Section: Introductionmentioning

confidence: 99%

An improved incremental least-mean-squares algorithm for distributed estimation over wireless sensor networks

Tan

Rong

et al. 2017

International Journal of Distributed Sensor Networks

View full text Add to dashboard Cite

Distributed parameter estimation problem has attracted much attention due to its important foundation of some applications in wireless sensor networks. In this work, we first investigate the performance tradeoff between existing incremental least-mean-squares algorithm and traditional steepest-descent algorithm from the aspects of initial convergence rate, steady-state convergence behavior, and oscillation, and present the motivations for improvements. Thereby we propose an improved incremental least-mean-squares distributed estimation algorithm that starts from the incremental least-mean-squares algorithm and works toward the objective of improvement on initial convergence rate and steadystate performance. The proposed algorithm meets the requirements of diminishing step size for wireless sensor networks without any increase in communication overhead and the mean stability condition is derived for a practical guideline. A target localization model in wireless sensor networks is used to illustrate the effect of the proposed algorithm. Simulation results confirm the performance improvement of our method, as well as the effectiveness of application in wireless sensor networks.

show abstract

Steady-State Analysis of the Deficient Length Incremental LMS Adaptive Networks

Cited by 13 publications

References 22 publications

Cooperative and distributed algorithm for compressed sensing recovery in WSNs

Cooperative and distributed algorithm for compressed sensing recovery in WSNs

Performance analysis of the deficient length NSAF algorithm and a variable step size method for improving its performance

An improved incremental least-mean-squares algorithm for distributed estimation over wireless sensor networks

Contact Info

Product

Resources

About