A Novel Family of Robust Incremental Adaptive Algorithms for Distributed Estimation Based on Bregman Divergence
Distributed estimation in the wireless sensor network (WSN) faces severe challenges when outliers are present in the desired data. Conventional estimation techniques such as the block-least-mean-square (BLMS) algorithm have been designed to perform well in the presence of Gaussian noise. However, the performance of these conventional techniques degrades significantly in the presence of outliers in the desired data. In this letter, a novel family of incremental distributed estimation techniques for WSN based on the Bregman divergence measure is proposed, which are named as incremental Kullback–Leibler-divergence-based BLMS algorithm, incremental Itakura–Saito-divergence-based BLMS algorithm, incremental Amari-alpha-divergence-based BLMS algorithm, and incremental beta-divergence-based BLMS algorithm. The local stability analysis shows the upper bound of the step size for the tractable analysis. Simulation results show that the proposed Bregman-divergence-based incremental algorithms perform better compared to the incremental BLMS algorithm in terms of the mean-square deviation.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Distributed adaptive algorithm based on the asymmetric cost of error functions
In this paper, a family of novel diffusion adaptive estimation algorithm is proposed from the asymmetric cost function perspective by combining diffusion strategy and the linear-linear cost (LLC), quadratic-quadratic cos…
Tracking Performance of Incremental LMS Algorithm over Adaptive Distributed Sensor Networks
In this paper we focus on the tracking performance of incremental adaptive LMS algorithm in an adaptive network. For this reason we consider the unknown weight vector to be a time varying sequence. First we analyze the p…
Distributed Adaptive Networks: A Graphical Evolutionary Game-Theoretic View
Distributed adaptive filtering has been considered as an effective approach for data processing and estimation over distributed networks. Most existing distributed adaptive filtering algorithms focus on designing differe…
Randomized Distributed Mean Estimation: Accuracy vs Communication
We consider the problem of estimating the arithmetic average of a finite collection of real vectors stored in a distributed fashion across several compute nodes subject to a communication budget constraint. Our analysis …
Fast Cross-Validation for Incremental Learning
Cross-validation (CV) is one of the main tools for performance estimation and parameter tuning in machine learning. The general recipe for computing CV estimate is to run a learning algorithm separately for each CV fold,…
Incremental Learning