Tesi etd-05092013-105913 |
Link copiato negli appunti
Tipo di tesi
Tesi di dottorato di ricerca
Autore
BORGESE, GIANLUCA
URN
etd-05092013-105913
Titolo
Design of embedded digital systems for data handling and elaboration in industrial and scientific applications
Settore scientifico disciplinare
ING-INF/01
Corso di studi
INGEGNERIA
Relatori
relatore Prof. Pace, Calogero
tutor Prof. Fanucci, Luca
tutor Prof. Fanucci, Luca
Parole chiave
- calcolo veloce
- emulazione
- FPGA
- protocolli d trasmissione
- sistemi embedded
- sistemi inerziali
Data inizio appello
24/06/2013
Consultabilità
Completa
Riassunto
Nowadays, in several research fields, both industrial and scientific, fast and
efficient data handling and elaboration is more and more important, especially in
view of the more and more growing expansion of systems that manage in real-time
a large amount of multi-media data. With data handling and elaboration it is
possible to define an big number of data manipulations (arithmetic-logic calculation,
compression, numerical filtering, storing, transmission, etc) executed by one or
more hardware/software systems within different applications. In the first phase of
this thesis work the concepts at the bottom of data handling and elaboration will be
introduced and the various typologies of "Embedded" systems will be discussed,
focusing mainly over the typologies treated in this thesis work. Subsequently, the
three embedded platforms designed will be treated in detail: the inertial/GPS
platform, based on two 8-bit microcontrollers, for the acquisition of data coming
from many sensors (thermometers, accelerometers and gyroscopes) and for the
management of other devices (GPS module and ZigBee module); the emulation
FPGA-based platform to conduct functional test on serial communication
interfaces, based on the new FF-LYNX protocol, modelized firstly in System-C and
then implemented on FPGA, to be used within high energy physics experiments;
the FPGA-based calculation platform. named DCMARK, that uses a CNN (Cellular
Neural Network) approach to rapidly solve non-linear differential equation at the
bottom of complex dynamical systems. All these works brought to interesting
scientific results within the design of embedded devices. In particular, the
inertial/GPS system demonstrated to be an efficient platform usable in different
fields such as body motion recognition, fall detection, aerial photogrammetry,
inertial navigation, etc and new ideas may be born after the deep integration of
system; the emulation system which allowed to validate and verify the proper
working of communication interfaces of FF-LYNX protocol, proving to be an
investigation instrument faster and more reliable than high level simulation; the
DCMARK calculation system which, using an innovative multi-processor approach,
allowed to tackle the solving of non-linear differential equation up to ten times more
quickly than the more modern multi-core processors.
efficient data handling and elaboration is more and more important, especially in
view of the more and more growing expansion of systems that manage in real-time
a large amount of multi-media data. With data handling and elaboration it is
possible to define an big number of data manipulations (arithmetic-logic calculation,
compression, numerical filtering, storing, transmission, etc) executed by one or
more hardware/software systems within different applications. In the first phase of
this thesis work the concepts at the bottom of data handling and elaboration will be
introduced and the various typologies of "Embedded" systems will be discussed,
focusing mainly over the typologies treated in this thesis work. Subsequently, the
three embedded platforms designed will be treated in detail: the inertial/GPS
platform, based on two 8-bit microcontrollers, for the acquisition of data coming
from many sensors (thermometers, accelerometers and gyroscopes) and for the
management of other devices (GPS module and ZigBee module); the emulation
FPGA-based platform to conduct functional test on serial communication
interfaces, based on the new FF-LYNX protocol, modelized firstly in System-C and
then implemented on FPGA, to be used within high energy physics experiments;
the FPGA-based calculation platform. named DCMARK, that uses a CNN (Cellular
Neural Network) approach to rapidly solve non-linear differential equation at the
bottom of complex dynamical systems. All these works brought to interesting
scientific results within the design of embedded devices. In particular, the
inertial/GPS system demonstrated to be an efficient platform usable in different
fields such as body motion recognition, fall detection, aerial photogrammetry,
inertial navigation, etc and new ideas may be born after the deep integration of
system; the emulation system which allowed to validate and verify the proper
working of communication interfaces of FF-LYNX protocol, proving to be an
investigation instrument faster and more reliable than high level simulation; the
DCMARK calculation system which, using an innovative multi-processor approach,
allowed to tackle the solving of non-linear differential equation up to ten times more
quickly than the more modern multi-core processors.
File
Nome file | Dimensione |
---|---|
Ph.D_The...RGESE.pdf | 6.41 Mb |
Contatta l’autore |