Catálogo de publicaciones - tesis

Compartir en
redes sociales

Título de Acceso Abierto

Métodos guiados por los datos para el análisis de señales: contribuciones a la descomposición empírica en modos

Marcelo Alejandro Colominas Gastón Schlotthauer Marcelo Risk Martín Hurtado Leandro Vignolo María Eugenia Torres

acceptedVersion.

Resumen/Descripción – provisto por el repositorio digital
Multicomponent signals, which are made of a superposition of a relatively small number of components with unique instantaneous frequencies, are useful to model a great number or real world signals, such as audio signals, biological signals o economic data series. The traditional analysis of these signals is performed through time-frequency or time-scale representations. Each component generates a ridge on the representation, which can be detected and, if desired, inversely transformed. According to this paradigm, one has one component for each detected ridge. This approach, of course, is not the only possible. The empirical mode decomposition (EMD) is a completely data-driven technique, which separates a signal into locally slow and fast oscillations. At the end, the original signal can be expressed as the sum a small number of modes, which can be represented in certain cases as AM-FM functions. However, since it is not based on time-frequency or time-scale representation, and therefore it does not identify one component for each ridge, but it can aggregate several ridges in one component, then it is capable of representing modes which are more complex than simple circular functions, becoming a more versatile method. An important limitation of EMD is that it is defined as an algorithm output, without solid theoretical bases. Because of this, and as a first contribution, we propose to see EMD as the solution of an unconstrained optimization problem as a first step to give this technique theoretical bases. We evaluate our proposal on artificial.
Palabras clave – provistas por el repositorio digital

Empirical mode decomposition; Time-frequency analysis; Time-scale analysis; Unconstrained optimization; Data-driven methods; Noise-assisted methods; Descomposición empírica en modos; Análisis tiempo-frecuencia; Análisis tiempo-escala; Optimización sin restricciones; Métodos guiados por los datos; Métodos asistidos por ruido

Disponibilidad
Institución detectada Año de publicación Navegá Descargá Solicitá
No requiere 2016 Biblioteca Virtual de la Universidad Nacional del Litoral (SNRD) acceso abierto

Información

Tipo de recurso:

tesis

Idiomas de la publicación

  • español castellano

País de edición

Argentina

Fecha de publicación