Publication: Diseño e implementación en FPGA de un filtro de partículas para aplicaciones biomédicas
Loading...
Official URL
Full text at PDC
Publication Date
2016
Authors
Advisors (or tutors)
Garnica Alcázar, Óscar
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
La diabetes mellitus tipo 1 (DM1) es una enfermedad crónica caracterizada por la incapacidad del páncreas de producir insulina. Esta hormona regula la absorción de la glucosa del torrente sanguíneo por parte de las células. Debido a la ausencia de insulina en el cuerpo, la glucosa se acumula en el torrente sanguíneo provocando problemas a corto y largo plazo, como por ejemplo deterioro celular. Los pacientes con esta enfermedad necesitan controlar su glucemia (concentración de glucosa en sangre) midiendo la misma de forma regular e inyectándose insulina subcutánea de por vida. Para conocer la glucemia se pueden utilizar Monitores Continuos de Glucosa (MCG), que proporcionan el valor de la glucosa intersticial en un rango entre uno y cinco minutos. Los MCG actuales presentan los siguientes problemas: Por un lado, el sensor que lleva incorporado introduce ruidos asociados a la medición obtenida. Y, por otro lado, el sensor se degrada a lo largo de su vida útil, lo que dificulta la interpretación de los datos obtenidos. La solución propuesta en este trabajo consiste en la utilización de filtros de partículas. Este tipo de filtros consta de cuatro fases: inicialización, predicción, corrección y remuestreo. Son capaces de identificar los estados ocultos del sistema (glucosa en sangre y degeneración del sensor), a partir de medidas indirectas del mismo (como por ejemplo la glucosa intersticial) teniendo en cuenta el ruido de las mediciones del MCG. En este proyecto se va a aplicar un filtro de partículas de cuatro estados (glucosa, velocidad de variación de la glucosa, degeneración del sensor y velocidad de variación de la degeneración del sensor.). En primera instancia, se utilizará la herramienta Matlab para analizar el correcto funcionamiento de este algoritmo frente a los problemas mencionados anteriormente de los MCG. Y, en segundo lugar, se realizará una implementación hardware sobre una FPGA.
Mellitus Diabetes type 1 (MD1) is a chronic disease characterized by the pancreas inability to produce insulin. This hormone regulates the absorption of glucose from the bloodstream by cells. Due to the absence of insulin in the body, glucose builds up in the bloodstream causing problems in short and long term, such as cellular deterioration. Patients with this disease need to control their glycemic (blood glucose concentration) by measuring it regularly and injecting subcutaneous insulin for life. In order to know glycemic can be used Continuous Glucose Monitor (CGM) which provide the value of interstitial glucose in a range between one and five minutes. Current CGM have the following problems: On one hand, the incorporated sensor introduces noise associated with the measurement obtained. And, on the other hand, the sensor degrades over its useful life, which makes data interpretation more difficult. The solution proposed in this paper is the use of particle filters. This type of filter consists in four phases: initialization, prediction, correction and resampling. They are able to identify the hidden states of the system (blood glucose and sensor degeneration) from indirect measurements thereof (interstitial glucose) considering CGM noise measurements. This project will apply a four states particle filter (glucose, glucose rate of change, sensor degeneration, and sensor degeneration rate of change). In the first instance, the Matlab tool will be used to analyze the correct functioning of this algorithm against the aforementioned problems of the CGM. And, second, a hardware implementation on an FPGA will be made.
Mellitus Diabetes type 1 (MD1) is a chronic disease characterized by the pancreas inability to produce insulin. This hormone regulates the absorption of glucose from the bloodstream by cells. Due to the absence of insulin in the body, glucose builds up in the bloodstream causing problems in short and long term, such as cellular deterioration. Patients with this disease need to control their glycemic (blood glucose concentration) by measuring it regularly and injecting subcutaneous insulin for life. In order to know glycemic can be used Continuous Glucose Monitor (CGM) which provide the value of interstitial glucose in a range between one and five minutes. Current CGM have the following problems: On one hand, the incorporated sensor introduces noise associated with the measurement obtained. And, on the other hand, the sensor degrades over its useful life, which makes data interpretation more difficult. The solution proposed in this paper is the use of particle filters. This type of filter consists in four phases: initialization, prediction, correction and resampling. They are able to identify the hidden states of the system (blood glucose and sensor degeneration) from indirect measurements thereof (interstitial glucose) considering CGM noise measurements. This project will apply a four states particle filter (glucose, glucose rate of change, sensor degeneration, and sensor degeneration rate of change). In the first instance, the Matlab tool will be used to analyze the correct functioning of this algorithm against the aforementioned problems of the CGM. And, second, a hardware implementation on an FPGA will be made.
Description
Máster en Ingeniería Informática, Facultad de Informática, Departamento de Arquitectura de Computadores y Automática, curso 2015-2016