Reports de recerca
http://hdl.handle.net/2117/3748
2017-02-21T16:57:28ZAprendizaje y asistencia virtual en red : la prueba Piloto : Cátedra Telefófica UPC : Análisis de le evolución y tendencias futuras de la sociedad de la información
http://hdl.handle.net/2117/86752
Aprendizaje y asistencia virtual en red : la prueba Piloto : Cátedra Telefófica UPC : Análisis de le evolución y tendencias futuras de la sociedad de la información
Fuentes Fort, Maria; González Bermúdez, Meritxell; Guardiola Garcia, Marta; Jofre Roca, Lluís; Romeu Robert, Jordi; Vallverdú Bayés, Francesc
Hemos llevado a cabo una prueba piloto, demostradora del uso de las tecnologías del lenguaje y el habla aplicadas al aprendizaje de inglés. En general, se espera que la actividad acerque al alumno el máximo posible a la realidad (simulación de entornos reales), donde se integren diferentes herramientas existentes en una sola para ofrecer al profesorado y al alumnado el valor añadido de la interacción, el feedback (necesario para la evaluación) y la reutilización de recursos y materiales existentes.
2016-05-09T09:28:23ZFuentes Fort, MariaGonzález Bermúdez, MeritxellGuardiola Garcia, MartaJofre Roca, LluísRomeu Robert, JordiVallverdú Bayés, FrancescHemos llevado a cabo una prueba piloto, demostradora del uso de las tecnologías del lenguaje y el habla aplicadas al aprendizaje de inglés. En general, se espera que la actividad acerque al alumno el máximo posible a la realidad (simulación de entornos reales), donde se integren diferentes herramientas existentes en una sola para ofrecer al profesorado y al alumnado el valor añadido de la interacción, el feedback (necesario para la evaluación) y la reutilización de recursos y materiales existentes.Internal report. Numerical methods for medium scale traveling ionospheric disturbances signal pre-processing
http://hdl.handle.net/2117/26832
Internal report. Numerical methods for medium scale traveling ionospheric disturbances signal pre-processing
Yang, Heng; Monte Moreno, Enrique; Hernández Pajares, Manuel
In this paper we discuss the e ects of several signal preprocessing
methods for enhancing the Medium Scale Traveling Ionospheric Dis-
turbances (MSTIDs) signal. The MSTIDs signal are the ionospheric
signatures of waves with a typical scale variation from 50 to 300 m/s,
which can be detected and modeled from variation of the ionospheric
Vertical Total Electron Content (VTEC) with dual-frequency mea-
surements of Global Navigation Satellite Systems (GNSS). In order to
enhance the useful information of the signal, and also reduce the noise,
we purpose the use of di erent numerical methods for preprocessing
the MSTIDs signal. A rst approach to preprocessing the signal is
a simple high-pass ltering or Slant Total Electron Content (STEC)
detrending of the ionospheric carrier phase LI [see Hernndez-Pajares
et al. 2006, Hernndez-Pajares et al. 2012]. In this work, we propose
the following signal processing steps in order to enhance the signal:
(1) Parabolic subtraction adopted in eliminating the typical geometric
length variation within the ionosphere carrier phase LI, (2) decima-
tion techniques used to instead of classic STEC detrending method,
(3) low pass ltering adapted to the properties of the desired signal.
From the results by testing the observing data on the day 1, 2011
obtained from the small area GNSS network in California, it's shown
that the MSTIDs signal after preprocessing techniques are more clear,
smooth and robust.
2015-03-19T11:17:45ZYang, HengMonte Moreno, EnriqueHernández Pajares, ManuelIn this paper we discuss the e ects of several signal preprocessing
methods for enhancing the Medium Scale Traveling Ionospheric Dis-
turbances (MSTIDs) signal. The MSTIDs signal are the ionospheric
signatures of waves with a typical scale variation from 50 to 300 m/s,
which can be detected and modeled from variation of the ionospheric
Vertical Total Electron Content (VTEC) with dual-frequency mea-
surements of Global Navigation Satellite Systems (GNSS). In order to
enhance the useful information of the signal, and also reduce the noise,
we purpose the use of di erent numerical methods for preprocessing
the MSTIDs signal. A rst approach to preprocessing the signal is
a simple high-pass ltering or Slant Total Electron Content (STEC)
detrending of the ionospheric carrier phase LI [see Hernndez-Pajares
et al. 2006, Hernndez-Pajares et al. 2012]. In this work, we propose
the following signal processing steps in order to enhance the signal:
(1) Parabolic subtraction adopted in eliminating the typical geometric
length variation within the ionosphere carrier phase LI, (2) decima-
tion techniques used to instead of classic STEC detrending method,
(3) low pass ltering adapted to the properties of the desired signal.
From the results by testing the observing data on the day 1, 2011
obtained from the small area GNSS network in California, it's shown
that the MSTIDs signal after preprocessing techniques are more clear,
smooth and robust.Internal report. Solving mixed integer non-linear programming problem applied to GNSS data
http://hdl.handle.net/2117/26831
Internal report. Solving mixed integer non-linear programming problem applied to GNSS data
Yang, Heng; Monte Moreno, Enrique; Hernández Pajares, Manuel
The purpose of this paper is to characterize Medium Scale Trav-
eling Ionospheric Disturbances (MSTIDs), by means of Mixed Integer
Nonlinear Programming (MINLP). The MINLP techniques are used
to for estimating the parameters of the equations that describe the
MSTIDs from a set of observations. A new MSTIDs wave detecting
method, which we will denote as Ambiguity Resolution in Global Navi-
gational Satellite System (GNSS) Ionospheric Interferometry (ARGII)
technique, is designed to model the MSTIDs wave with the data from
the wide low-density GNSS receivers network. The ARGII techniques
can be set as an special instance of MINLP, because the problem is set
as a series of MSTIDs equations including the unknown wave veloc-
ity (continuous) and cycle ambiguities (integers). The performance of
heuristic and direct search optimization algorithms are evaluated by
solving the MINLP problem with techniques bared an di erent prin-
ciples, and as benchmark we use the solution obtained by exhaustive
enumeration of all possible integer solutions. Among the algorithms
we have implemented in this work are genetic algorithm, simulated
annealing, particle swarm, pattern search and Nelder Mead methods.
The GNSS data used to test the these solvers is observed from the wide
GNSS network in the north of Poland on the day 353, 2013 whose di-
ameter is more than the half of wavelength and therefore will have
phase ambiguities. The evaluating experiments show that the results
computed by the simple improved optimization algorithms especially
the Nelder Mead have not only high correlations with the reference
method (i.e. exhaustive enumeration) but also extremely lower time
complexity compared to the benchmark method. Despite unguaran-
teed global optimal results for the MINLP problems, these methods
show the excellent performance in time complexity when computing
the velocities of MSTIDs with ARGII techniques from large quantity
of the GNSS data.
Internal Report
2015-03-19T11:14:49ZYang, HengMonte Moreno, EnriqueHernández Pajares, ManuelThe purpose of this paper is to characterize Medium Scale Trav-
eling Ionospheric Disturbances (MSTIDs), by means of Mixed Integer
Nonlinear Programming (MINLP). The MINLP techniques are used
to for estimating the parameters of the equations that describe the
MSTIDs from a set of observations. A new MSTIDs wave detecting
method, which we will denote as Ambiguity Resolution in Global Navi-
gational Satellite System (GNSS) Ionospheric Interferometry (ARGII)
technique, is designed to model the MSTIDs wave with the data from
the wide low-density GNSS receivers network. The ARGII techniques
can be set as an special instance of MINLP, because the problem is set
as a series of MSTIDs equations including the unknown wave veloc-
ity (continuous) and cycle ambiguities (integers). The performance of
heuristic and direct search optimization algorithms are evaluated by
solving the MINLP problem with techniques bared an di erent prin-
ciples, and as benchmark we use the solution obtained by exhaustive
enumeration of all possible integer solutions. Among the algorithms
we have implemented in this work are genetic algorithm, simulated
annealing, particle swarm, pattern search and Nelder Mead methods.
The GNSS data used to test the these solvers is observed from the wide
GNSS network in the north of Poland on the day 353, 2013 whose di-
ameter is more than the half of wavelength and therefore will have
phase ambiguities. The evaluating experiments show that the results
computed by the simple improved optimization algorithms especially
the Nelder Mead have not only high correlations with the reference
method (i.e. exhaustive enumeration) but also extremely lower time
complexity compared to the benchmark method. Despite unguaran-
teed global optimal results for the MINLP problems, these methods
show the excellent performance in time complexity when computing
the velocities of MSTIDs with ARGII techniques from large quantity
of the GNSS data.Internal report. Spectral methods for time delay estimation of MSTIDs signal
http://hdl.handle.net/2117/26830
Internal report. Spectral methods for time delay estimation of MSTIDs signal
Yang, Heng; Monte Moreno, Enrique; Hernández Pajares, Manuel
The primary goal of this work is to summarize spectral meth-
ods to estimate the time delay variations of the detrended Medium
Scale Traveling Ionospheric Disturbances (MSTIDs) signals and pre-
liminarily discuss the potential infuence of the Doppler e fects. The
MSTIDs signals are assumed as the planar waves with the wavelength
between 10-100 kilometers, which are measured from a local high-
dense receivers Global Navigation Satellite System (GNSS) network
for a given satellite. The signals show strong correlation with each
others in the whole observation time series. The time delay of the
detrended MSTIDs signals preprocessed with a speci c overlapped
sliding window can be estimated by several techniques. The data con-
sisted on two detrended MSTIDs signals measured from the GNSS
data on the day 1, 2011 obtained from the small area network in Cal-
ifornia. In this paper we have implemented the following methods
in order to estimate the time delay between signals: (1) Generalized
cross correlation method, which estimates the time delay in time do-
main after the inverse Fourier transform of appropriately weighted
cross power spectrum density, (2) generalized phase spectrum method
directly measured from the appropriately weighted cross power spec-
trum density, (3) phase spectrum di erence method calculated ap-
proximately from the fundamental waves of the signals. The results
with di erent methods show the similar variation of time delay of two
MSTIDs signals. In addition, with the phase spectrum method, the
results can present the clear frequency variation of MSTIDs signals
which may be in
uenced by potential Doppler e ect. Note that the
correct estimate of the delays required of a preprocessing of the data.
The rst di erence of the Vertical Total Electron Contents (dVTEC)
signal was preprocessed by modeling the e ect of the displacement of
the satellite by means of a parabolic component, which is subtracted,
following by detrend of 60-second intervals, decimation and a further
low pass ltering process.
Internal Report
2015-03-19T11:11:03ZYang, HengMonte Moreno, EnriqueHernández Pajares, ManuelThe primary goal of this work is to summarize spectral meth-
ods to estimate the time delay variations of the detrended Medium
Scale Traveling Ionospheric Disturbances (MSTIDs) signals and pre-
liminarily discuss the potential infuence of the Doppler e fects. The
MSTIDs signals are assumed as the planar waves with the wavelength
between 10-100 kilometers, which are measured from a local high-
dense receivers Global Navigation Satellite System (GNSS) network
for a given satellite. The signals show strong correlation with each
others in the whole observation time series. The time delay of the
detrended MSTIDs signals preprocessed with a speci c overlapped
sliding window can be estimated by several techniques. The data con-
sisted on two detrended MSTIDs signals measured from the GNSS
data on the day 1, 2011 obtained from the small area network in Cal-
ifornia. In this paper we have implemented the following methods
in order to estimate the time delay between signals: (1) Generalized
cross correlation method, which estimates the time delay in time do-
main after the inverse Fourier transform of appropriately weighted
cross power spectrum density, (2) generalized phase spectrum method
directly measured from the appropriately weighted cross power spec-
trum density, (3) phase spectrum di erence method calculated ap-
proximately from the fundamental waves of the signals. The results
with di erent methods show the similar variation of time delay of two
MSTIDs signals. In addition, with the phase spectrum method, the
results can present the clear frequency variation of MSTIDs signals
which may be in
uenced by potential Doppler e ect. Note that the
correct estimate of the delays required of a preprocessing of the data.
The rst di erence of the Vertical Total Electron Contents (dVTEC)
signal was preprocessed by modeling the e ect of the displacement of
the satellite by means of a parabolic component, which is subtracted,
following by detrend of 60-second intervals, decimation and a further
low pass ltering process.Extended version. Occurrence of solar flares viewed with GPS: statistics and fractal nature
http://hdl.handle.net/2117/24607
Extended version. Occurrence of solar flares viewed with GPS: statistics and fractal nature
Monte Moreno, Enrique; Hernández Pajares, Manuel
In this paper we describe the statistical properties of the EUV solar flux sudden variation. The solar flux variation is modeled as a time series characterized by the subsolar VTEC (Vertical Total Electron Content) double-difference in time, computed with dual frequency GNSS (Global Navigation Satellite System) measurements in the daylight hemisphere. By assuming a sudden overionization pattern of solar origin, during the last solar cycle, we propose a model that explains it's characteristics, and the forecasting limitations. The two defining characteristics of this time series, is an extreme variability (i.e.\ in a solar cycle one can find events at $400 \sigma$ from the mean value) and a temporal correlation that is independent of the time scale. We give a characterization of a model that explains the empirical results, and properties such as, a) the persistence and presence of bursts of solar flares, b) their long tail peak values of the solar flux variation. We show that the solar flux variation time series can be characterized by a fractional Brownian model for the long term dependence, and a powerlaw distribution for the extreme values that appear in the time series.
Extended version of the paper
2014-11-07T14:46:29ZMonte Moreno, EnriqueHernández Pajares, ManuelIn this paper we describe the statistical properties of the EUV solar flux sudden variation. The solar flux variation is modeled as a time series characterized by the subsolar VTEC (Vertical Total Electron Content) double-difference in time, computed with dual frequency GNSS (Global Navigation Satellite System) measurements in the daylight hemisphere. By assuming a sudden overionization pattern of solar origin, during the last solar cycle, we propose a model that explains it's characteristics, and the forecasting limitations. The two defining characteristics of this time series, is an extreme variability (i.e.\ in a solar cycle one can find events at $400 \sigma$ from the mean value) and a temporal correlation that is independent of the time scale. We give a characterization of a model that explains the empirical results, and properties such as, a) the persistence and presence of bursts of solar flares, b) their long tail peak values of the solar flux variation. We show that the solar flux variation time series can be characterized by a fractional Brownian model for the long term dependence, and a powerlaw distribution for the extreme values that appear in the time series.A multivariate neural network approach to tourism demand forecasting
http://hdl.handle.net/2117/23086
A multivariate neural network approach to tourism demand forecasting
Claveria, Oscar; Monte Moreno, Enrique; Torra, Salvador
This study compares the performance of different Artificial Neural Networks models for tourist demand forecasting in a multiple-output framework. We test the forecasting accuracy of three different types of architectures: a multi-layer perceptron network, a radial basis function network and an Elman neural network. We use official statistical data of inbound international tourism demand to Catalonia (Spain) from 2001 to 2012. By means of cointegration analysis we find that growth rates of tourist arrivals from all different countries share a common stochastic trend, which leads us to apply a multivariate out-of-sample forecasting comparison. When comparing the forecasting accuracy of the different techniques for each visitor market and for different forecasting horizons, we find that radial basis function models outperform multi-layer perceptron and Elman networks. We repeat the experiment assuming different topologies regarding the number of lags used for concatenation so as to evaluate the effect of the memory on the forecasting results, and we find no significant differences when additional lags are incorporated. These results reveal the suitability of hybrid models such as radial basis functions that combine supervised and unsupervised learning for economic forecasting with seasonal data.
2014-05-28T14:30:55ZClaveria, OscarMonte Moreno, EnriqueTorra, SalvadorThis study compares the performance of different Artificial Neural Networks models for tourist demand forecasting in a multiple-output framework. We test the forecasting accuracy of three different types of architectures: a multi-layer perceptron network, a radial basis function network and an Elman neural network. We use official statistical data of inbound international tourism demand to Catalonia (Spain) from 2001 to 2012. By means of cointegration analysis we find that growth rates of tourist arrivals from all different countries share a common stochastic trend, which leads us to apply a multivariate out-of-sample forecasting comparison. When comparing the forecasting accuracy of the different techniques for each visitor market and for different forecasting horizons, we find that radial basis function models outperform multi-layer perceptron and Elman networks. We repeat the experiment assuming different topologies regarding the number of lags used for concatenation so as to evaluate the effect of the memory on the forecasting results, and we find no significant differences when additional lags are incorporated. These results reveal the suitability of hybrid models such as radial basis functions that combine supervised and unsupervised learning for economic forecasting with seasonal data.Informe proyecto SARAI
http://hdl.handle.net/2117/22261
Informe proyecto SARAI
Hernando Pericás, Francisco Javier
Informe de actividades en el proyecto SARAI durante 2012
2014-03-18T15:13:14ZHernando Pericás, Francisco JavierCorpus selection
http://hdl.handle.net/2117/21703
Corpus selection
Adda, Gilles; Barras, Claude; Kernal Ekenel, Hazim; Morros Rubió, Josep Ramon; Hernando Pericás, Francisco Javier
Entregable del proyecto Collaborative Annotation of multi-MOdal, MultI-Lingual and multi-mEdia documents. This document describes the different corpora that will be used during the Camomile project
2014-02-21T18:55:37ZAdda, GillesBarras, ClaudeKernal Ekenel, HazimMorros Rubió, Josep RamonHernando Pericás, Francisco JavierEntregable del proyecto Collaborative Annotation of multi-MOdal, MultI-Lingual and multi-mEdia documents. This document describes the different corpora that will be used during the Camomile projectReport on awareness, mobilisation and dissemination actions
http://hdl.handle.net/2117/21110
Report on awareness, mobilisation and dissemination actions
Trandaba¿, Diana; Cristea, Dan; Branco, Antonio; Mendes, Amalia; Pellegrini, Thomas; Ananiadou, Sophia; Thompson, Paul; Tufis, Dan; Gilmenau, Georgiana; Rosner, Mike; Moreno Bilbao, M. Asunción; Bel, Nuria
The central objective of the Metanet4u project is to contribute to the establishment of a pan-European digital platform that makes available language resources and services, encompassing both datasets and software tools, for speech and language processing, and supports a new generation of exchange facilities for them.
2013-12-31T10:20:47ZTrandaba¿, DianaCristea, DanBranco, AntonioMendes, AmaliaPellegrini, ThomasAnaniadou, SophiaThompson, PaulTufis, DanGilmenau, GeorgianaRosner, MikeMoreno Bilbao, M. AsunciónBel, NuriaThe central objective of the Metanet4u project is to contribute to the establishment of a pan-European digital platform that makes available language resources and services, encompassing both datasets and software tools, for speech and language processing, and supports a new generation of exchange facilities for them.Sustainability strategy and plans beyond the end of the project
http://hdl.handle.net/2117/21058
Sustainability strategy and plans beyond the end of the project
Trandaba¿, Diana; Cristea, Dan; Branco, Antonio; Mendes, Amalia; Pellegrini, Thomas; Ananiadou, Sophia; Thompson, Paul; Tufis, Dan; Gilmenau, Georgiana; Rosner, Mike; Moreno Bilbao, M. Asunción; Bel, Nuria
The central objective of the Metanet4u project is to contribute to the establishment of a pan-European digital platform that makes available language resources and services, encompassing both datasets and software tools, for speech and language processing, and supports a new generation of exchange facilities for them.
2013-12-18T18:42:47ZTrandaba¿, DianaCristea, DanBranco, AntonioMendes, AmaliaPellegrini, ThomasAnaniadou, SophiaThompson, PaulTufis, DanGilmenau, GeorgianaRosner, MikeMoreno Bilbao, M. AsunciónBel, NuriaThe central objective of the Metanet4u project is to contribute to the establishment of a pan-European digital platform that makes available language resources and services, encompassing both datasets and software tools, for speech and language processing, and supports a new generation of exchange facilities for them.