2011
Achutegui, Katrin; Miguez, Joaquin
A Parallel Resampling Scheme and its Application to Distributed Particle Filtering in Wireless Networks Proceedings Article
En: 2011 4th IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP), pp. 81–84, IEEE, San Juan, 2011, ISBN: 978-1-4577-2105-2.
Resumen | Enlaces | BibTeX | Etiquetas: Approximation algorithms, Approximation methods, Artificial neural networks, distributed resampling, DRNA technique, Markov processes, nonproportional allocation algorithm, parallel resampling scheme, PF, quantization, Signal processing, Vectors, Wireless sensor network, Wireless Sensor Networks, WSN
@inproceedings{Achutegui2011,
title = {A Parallel Resampling Scheme and its Application to Distributed Particle Filtering in Wireless Networks},
author = {Katrin Achutegui and Joaquin Miguez},
url = {http://ieeexplore.ieee.org/articleDetails.jsp?arnumber=6136051},
isbn = {978-1-4577-2105-2},
year = {2011},
date = {2011-01-01},
booktitle = {2011 4th IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP)},
pages = {81--84},
publisher = {IEEE},
address = {San Juan},
abstract = {We address the design of a particle filter (PF) that can be implemented in a distributed manner over a network of wireless sensor nodes, each of them collecting their own local data. This is a problem that has received considerable attention lately and several methods based on consensus, the transmission of likelihood information, the truncation and/or the quantization of data have been proposed. However, all existing schemes suffer from limitations related either to the amount of required communications among the nodes or the accuracy of the filter outputs. In this work we propose a novel distributed PF that is built around the distributed resampling with non-proportional allocation (DRNA) algorithm. This scheme guarantees the properness of the particle approximations produced by the filter and has been shown to be both efficient and accurate when compared with centralized PFs. The standard DRNA technique, however, places stringent demands on the communications among nodes that turn out impractical for a typical wireless sensor network (WSN). In this paper we investigate how to reduce this communication load by using (i) a random model for the spread of data over the WSN and (ii) methods that enable the out-of-sequence processing of sensor observations. A simple numerical illustration of the performance of the new algorithm compared with a centralized PF is provided.},
keywords = {Approximation algorithms, Approximation methods, Artificial neural networks, distributed resampling, DRNA technique, Markov processes, nonproportional allocation algorithm, parallel resampling scheme, PF, quantization, Signal processing, Vectors, Wireless sensor network, Wireless Sensor Networks, WSN},
pubstate = {published},
tppubtype = {inproceedings}
}
We address the design of a particle filter (PF) that can be implemented in a distributed manner over a network of wireless sensor nodes, each of them collecting their own local data. This is a problem that has received considerable attention lately and several methods based on consensus, the transmission of likelihood information, the truncation and/or the quantization of data have been proposed. However, all existing schemes suffer from limitations related either to the amount of required communications among the nodes or the accuracy of the filter outputs. In this work we propose a novel distributed PF that is built around the distributed resampling with non-proportional allocation (DRNA) algorithm. This scheme guarantees the properness of the particle approximations produced by the filter and has been shown to be both efficient and accurate when compared with centralized PFs. The standard DRNA technique, however, places stringent demands on the communications among nodes that turn out impractical for a typical wireless sensor network (WSN). In this paper we investigate how to reduce this communication load by using (i) a random model for the spread of data over the WSN and (ii) methods that enable the out-of-sequence processing of sensor observations. A simple numerical illustration of the performance of the new algorithm compared with a centralized PF is provided.
2009
Goez, Roger; Lazaro, Marcelino
Training of Neural Classifiers by Separating Distributions at the Hidden Layer Proceedings Article
En: 2009 IEEE International Workshop on Machine Learning for Signal Processing, pp. 1–6, IEEE, Grenoble, 2009, ISBN: 978-1-4244-4947-7.
Resumen | Enlaces | BibTeX | Etiquetas: Artificial neural networks, Bayesian methods, Cost function, Curve fitting, Databases, Function approximation, Neural networks, Speech recognition, Support vector machine classification, Support vector machines
@inproceedings{Goez2009,
title = {Training of Neural Classifiers by Separating Distributions at the Hidden Layer},
author = {Roger Goez and Marcelino Lazaro},
url = {http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=5306240},
isbn = {978-1-4244-4947-7},
year = {2009},
date = {2009-01-01},
booktitle = {2009 IEEE International Workshop on Machine Learning for Signal Processing},
pages = {1--6},
publisher = {IEEE},
address = {Grenoble},
abstract = {A new cost function for training of binary classifiers based on neural networks is proposed. This cost function aims at separating the distributions for patterns of each class at the output of the hidden layer of the network. It has been implemented in a Generalized Radial Basis Function (GRBF) network and its performance has been evaluated under three different databases, showing advantages with respect to the conventional Mean Squared Error (MSE) cost function. With respect to the Support Vector Machine (SVM) classifier, the proposed method has also advantages both in terms of performance and complexity.},
keywords = {Artificial neural networks, Bayesian methods, Cost function, Curve fitting, Databases, Function approximation, Neural networks, Speech recognition, Support vector machine classification, Support vector machines},
pubstate = {published},
tppubtype = {inproceedings}
}
A new cost function for training of binary classifiers based on neural networks is proposed. This cost function aims at separating the distributions for patterns of each class at the output of the hidden layer of the network. It has been implemented in a Generalized Radial Basis Function (GRBF) network and its performance has been evaluated under three different databases, showing advantages with respect to the conventional Mean Squared Error (MSE) cost function. With respect to the Support Vector Machine (SVM) classifier, the proposed method has also advantages both in terms of performance and complexity.