1359 research outputs found
Sort by
Consensus Algorithms for Estimation and Discrete Averaging in Networked Control Systems
Get PDFIn this thesis several topics on consensus and gossip algorithms for multi-agent systems are addressed. An agent is a dynamical system that can be fully described by a state-space representation of its dynamics. A multi-agent system is a network of agents whose pattern of interactions or couplings is described by a graph. Consensus problems in multi-agent systems consist in the study of local interaction rules between the agents such that as global emergent behavior the network converges to the so called "consensus" or "agreement" state where the value of each agent's state is the same and it is possibly a function of the initial network state, for instance the average. A consensus algorithm is thus a set of local interaction rules that solve the consensus problem under some assumptions on the network topology.
A gossip algorithm is a set of local state update rules between the agents that, disregarding their objective, are supposed to be implemented in a totally asynchronous way between pairs of neighboring agents, thus resembling the act of "gossiping" in a crowd of people.
In this thesis several algorithms based on gossip that solve the consensus and other related problems are presented.
In the �first part, several solutions to the consensus problem based on gossip under different sets of assumptions are proposed. In the fi�rst case, it is assumed that the state of the agents is discretized and represents a collection of tasks of different size. In the second case,
under the same discretization assumptions of the �rst case, it is assumed that the network is represented by a Hamiltonian graph and it is shown how under this assumption the convergence speed can be improved. In the third case, a solution for the consensus problem for networks represented by arbitrary strongly connected directed graphs is proposed, assuming that the state of the agents is a real number.
In the fourth case, a coordinate-free consensus algorithm based on gossip is designed and applied to a network of vehicles able to sense the relative distance between each other but with no access to absolute position information or to a common coordinate system. The proposed algorithm is then used to build in a decentralized way a common reference frame for the network of vehicles.
In the second part, a novel local interaction rule based on the consensus equation is proposed together with an algorithm to estimate in a decentralized way the spectrum of the Laplacian matrix that encodes the network topology. As emergent behavior, each agent's state oscillates only at frequencies corresponding to the eigenvalues of the
Laplacian matrix thus mapping the spectrum estimation problem into a signal processing problem solvable using the Fourier Transform. It is further shown that the constant component of the emergent behavior in the frequency domain solves the consensus on the average problem.
The spectrum estimation algorithm is then applied to leader-follower networks of mobile vehicles to infer in a decentralized way properties such as controllability, osservability and other topological features of
the network such as its topology.
Finally, a fault detection and recovery technique for sensor networks based on the so called motion-probes is presented to address the inherent lack of robustness against outlier agents in networks implementing
consensus algorithms to solve the distributed averaging problem
Aggiustamenti metabolici durante il recupero nell'Interval Training: studio in laboratorio e sul campo in atleti di mezzofondo e velocità prolungata
Get PDFLo scopo di questo studio è stato quello di quantificare l'affidabilità della frequenza cardiaca (HR), come parametro utile per valutare le condizioni di recupero di mezzofondisti e velocisti, durante un Interval Training (IT). Sono stati studiati ventidue atleti sottoposti a due diverse sessioni di
IT: al 80% e 120% della soglia anaerobica (SA). Durante entrambe le sessioni abbiamo misurato
HR, consumo di ossigeno (VO₂), produzione di anidride carbonica (VCO₂), ventilazione polmonare
(Ve), mediante un analizzatore portatile di gas. Inoltre, è stato calcolato l'eccesso di produzione di anidride carbonica (CO₂excess) per verificare l'ipotesi che la deriva cardiaca, che si è verificata durante il recupero tra le ripetizioni di IT, fosse collegata ad uno stato di acidosi lattica. La deriva cardiaca è avvenuta in entrambi i gruppi ad entrambe le intensità lavorative, mentre CO₂excess, VO₂ e Ve non si sono modificati. Sulla base della nostra ricerca, l'utilizzo della HR come parametro di valutazione del recupero metabolico appare impreciso e fuorviante ed in definitiva non idoneo per la gestione del recupero nell’allenamento.
----------------------------------------------------------------------------------------------------------------------------The aim of this study was to quantify the reliability of heart rate (HR) monitoring to assess
recovery status, after moderate and high-intensity exercise, in prolonged sprint and middle
distance running specialties of athletics. Twenty-two runners underwent two different sessions of
interval training: at 80% and 120% of anaerobic threshold (AT). Throughout both sessions we
measured HR, oxygen uptake (VO₂), carbon dioxide production (VCO₂), pulmonary ventilation
(Ve), by means a portable gas analyzer. The excess in carbon dioxide production (CO₂excess) was
also measured to test the hypothesis that a cardiac drift, that occurred during recovery between
repetitions of interval training, was related to lactic acidosis. The heart rate charts in both groups
of athletes and with respect to the two workloads showed a progressive increase of the same
during recovery periods, while CO2excess as well as VO₂ and Ve were not modified. On the basis
of our research, the use of HR as a parameter of evaluation of the established recovery is
inaccurate and unfit for training
Quantificazione dei fenomeni di arricchimento degli elementi in traccia lungo la catena trofica come strumento per la valutazione della qualità degli ecosistemi marini e delle risorse demersali
Get PDFEspecially in the last centuries (Barbante et al., 2004), several anthropogenic activities have led to an
increase of heavy metals global emissions (Gašpić et al., 2002; Fabris et al., 2006).
Among trace elements released in environment, some are essential for organisms (Rainbow, 1988;
Bustamante et al., 2000; Canli and Atli, 2003), due to their role in metabolic functioning (Barkay et al.,
1992). Others play an important role in detoxification mechanisms, whereas those defined as not
essential have no known role in biological system (Bustamante et al., 2000; Canli and Atli, 2003). Not
essential metals are toxic even at low concentrations, thus representing a risk for human health
(Barkay et al., 1992; Edwards et al., 2001).
Heavy metals accumulation by marine organisms is a well known phenomenon (White and Rainbow,
1982; Bacci and Gaggi, 1998; Sandor et al., 2001; Canli and Atli, 2003; Mishra et al., 2007), especially in
predator species at the top of the food chains (Storelli et al., 2001; Domi et al., 2005; Coelho et al.,
2008; Suk et al., 2009). Predator species are often economically important and directly consumed by
humans (Das et al., 2003; Debacker et al., 2001; Kai et al., 1978; Meaburn, 1978; Das et al., 2000;
Storelli et al., 2001).
This study’s aim was to provide data about cadmium, total mercury and lead concentrations in the
muscle of Merluccius merluccius and Scyliorhinus canicula, that may be considered as good bioindicators
of the extend of these elements contamination in the Sardinian waters.
The behavior of elements was assessed in function of several parameters, with particular attention to
the study of the bioaccumulation across trophic levels, estimated through stomach content analysis for
both species (Schetagne et al., 2000; Storelli and Marcotrigiano, 2004; Kasper et al., 2009). Moreover,
some hazard indices were calculated in order to measure potential risks for human race.
Several factors seemed to interact with feeding dynamics of both species, among these, growing is
probably the most influential. In addition, changes related to the reproductive process, like the
decrease of feeding intensity in M. merluccius or the different bathymetric distribution of the two
sexes in S. canicula, were known to be important as well. Finally, remarkable variations in the diet
were related to the availability and abundance of preys, both in terms of geographic distributions and
bathymetric distribution.
Highest concentration of total mercury were found in the muscle of S. canicula, while lead was slightly
more concentrated in M. merluccius. No differences for cadmium in both species were observed.Specie-specific differences were likely due to an higher intake rate or to a greater susceptibility to
heavy metals contamination characteristics of the Chondrichthyes, as previously reported by Jeffree et
al (2006) and Mathews et al. (2008).
From a regulatory point of view regarding cadmium and lead, the intake of both species could be
considered safe. On the contrary, health potential risks would arise from the intake of total and
methylmercury.
Consequently, despite the contribution of fish to diet is fundamental in maintaining a good state of
health, regular consumption of both species, especially S. canicula, as the high levels of mercury should
be discouraged.
Moreover, it is necessary to note that intakes estimated in this work could be underestimated. They, in
fact, do not take account for the intake from other food than fish, the variation in the consumption rate
observed in different geographic areas, or the consumption of different tissues (liver, kidney, gonads),
probably characterized by diverse concentrations of metals.
In conclusion, the output of this work has to be considered as preliminary. More wide-ranging studies
are necessary in order to better assess the environmental impact of heavy metals and the health risks
related to the intake of fish from Sardinian waters
Interactive search techniques for content-based retrieval from archives of images
Get PDFThrough a little investigation by file types it is possible to easily find that one of the most popular
search engines has in its indexes about 10 billion of images. Even considering that this
data is probably an underestimate of the real number, however, immediately it gives us an
idea of how the images are a key component in human communication. This so exorbitant
number puts us in the face of the enormous difficulties encountered when one has to deal
with them. Until now, the images have always been accompanied by textual data: description,
tags, labels, ... which are used to retrieve them fromthe archives. However it is clear that
their increase, occurred in recent years, does not allow this type cataloguing. Furthermore,
for its own nature, a manual cataloguing is subjective, partial and without doubt subject to
error. To overcome this situation in recent years it has gotten a footing a kind of search based
on the intrinsic characteristics of images such as colors and shapes. This information is then
converted into numerical vectors, and through their comparison it is possible to find images
that have similar characteristics. It is clear that a search, on this level of representation of the
images, is far from the user perception that of the images.
To allow the interaction between users and retrieval systems and improve the performance,
it has been decided to involve the user in the search allowing to him to give a feedback
of relevance of the images retrieved so far. In this the kind of image that are interesting
for user can be learnt by the system and an improvement in the next iteration can be obtained.
These techniques, although studied for many years, still present open issues. High
dimensional feature spaces, lack of relevant training images, and feature spaceswith lowdiscriminative
capability are just some of the problems encountered. In this thesis these problems
will be faced by proposing some innovative solutions both to improve performance
obtained by methods proposed in the literature, and to provide to retrieval systems greater
generalization capability. Techniques of data fusion, both at the feature space level and at
the level of different retrieval techniques, will be presented, showing that the former allow
greater discriminative capability while the latter provide more robustness to the system. To
overcome the lack of images of training it will be proposed a method to generate synthetic
patterns allowing in this way a more balanced learning. Finally, new methods to measure
similarity between images and to explore more efficiently the feature space will be proposed.
The presented results show that the proposed approaches are indeed helpful in resolving
some of the main problems in content based image retrieval
Real-time GPU-accelerated Out-of-Core Rendering and Light-field Display Visualization for Improved Massive Volume Understanding
Get PDFNowadays huge digital models are becoming increasingly available for a number of different applications ranging from CAD, industrial design to medicine and natural sciences. Particularly, in the field of medicine, data acquisition devices such as MRI or CT scanners routinely produce huge volumetric datasets. Currently, these datasets can easily reach dimensions of 1024^3 voxels and datasets larger than that are not uncommon.
This thesis focuses on efficient methods for the interactive exploration of such large volumes using direct volume visualization techniques on commodity platforms. To reach this goal specialized multi-resolution structures and algorithms, which are able to directly render volumes of potentially unlimited size are introduced. The developed techniques are output sensitive and their rendering costs depend only on the complexity of the generated images and not on the complexity of the input datasets. The advanced characteristics of modern GPGPU architectures are exploited and combined with an out-of-core framework in order to provide a more flexible, scalable and efficient implementation of these algorithms and data structures on single GPUs and GPU clusters.
To improve visual perception and understanding, the use of novel 3D display technology based on a light-field approach is introduced. This kind of device allows multiple naked-eye users to perceive virtual objects floating inside the display workspace, exploiting the stereo and horizontal parallax. A set of specialized and interactive illustrative techniques capable of providing different contextual information in different areas of the display, as well as an out-of-core CUDA based ray-casting engine with a number of improvements over current GPU volume ray-casters are both reported. The possibilities of the system are demonstrated by the multi-user interactive exploration of 64-GVoxel datasets on a 35-MPixel light-field display driven by a cluster of PCs. ------------------------------------------------------------------------------------------------------
Negli ultimi anni si sta verificando una proliferazione sempre più consistente di modelli
digitali di notevoli dimensioni in campi applicativi che variano dal CAD e la progettazione
industriale alla medicina e le scienze naturali. In modo particolare, nel settore della medicina,
le apparecchiature di acquisizione dei dati come RM o TAC producono comunemente dei
dataset volumetrici di grosse dimensioni. Questi dataset possono facilmente raggiungere
taglie dell’ordine di 10243 voxels e dataset di dimensioni maggiori possono essere frequenti.
Questa tesi si focalizza su metodi efficienti per l’esplorazione di tali grossi volumi utilizzando
tecniche di visualizzazione diretta su piattaforme HW di diffusione di massa. Per
raggiungere tale obiettivo si introducono strutture specializzate multi-risoluzione e algoritmi
in grado di visualizzare volumi di dimensioni potenzialmente infinite. Le tecniche sviluppate
sono “ouput sensitive” e la loro complessità di rendering dipende soltanto dalle dimensioni
delle immagini generate e non dalle dimensioni dei dataset di input. Le caratteristiche avanzate
delle architetture moderne GPGPU vengono inoltre sfruttate e combinate con un framework
“out-of-core” in modo da offrire una implementazione di questi algoritmi e strutture
dati più flessibile, scalabile ed efficiente su singole GPU o cluster di GPU.
Per migliorare la percezione visiva e la comprensione dei dati, viene introdotto inoltre l’uso
di tecnologie di display 3D di nuova generazione basate su un approccio di tipo light-field.
Questi tipi di dispositivi consentono a diversi utenti di percepire ad occhio nudo oggetti che
galleggiano all’interno dello spazio di lavoro del display, sfruttando lo stereo e la parallasse
orizzontale. Si descrivono infine un insieme di tecniche illustrative interattive in grado di
fornire diverse informazioni contestuali in diverse zone del display, così come un motore di
“ray-casting out-of-core” basato su CUDA e contenente una serie di miglioramenti rispetto
agli attuali metodi GPU di “ray-casting” di volumi. Le possibilità del sistema sono dimostrate
attraverso l’esplorazione interattiva di dataset di 64-GVoxel su un display di tipo light-field
da 35-MPixel pilotato da un cluster di PC
Composition control and estimation with temperature measurements for multicomponent distillation columns
Get PDFIn this Thesis, the problem of controlling and estimating compositions of interest in a multicomponent column with temperature measurements is addressed. This work is motivated by the necessity of (i) regulating some of the effluent compositions around the prescribed values for a given separation, in spite of disturbances entering the column as fluctuations in feed flow rate, composition and temperature, and (ii) estimating such compositions in order to monitor the separation.
These problems cannot be practically solved by composition analyzers, because of their high costs of purchase and maintenance, their reliability issues and the delays in obtaining composition measurements. A possible solution is the employment of temperature measurements, which do not present the abovementioned shortcomings. In a multicomponent distillation column, the compositions in a tray are not uniquely determined by the temperature at the same tray, and therefore temperature sensors cannot provide composition control and estimation without offset. Thus, a methodology that permits the achievement of composition control and estimation objectives within a predetermined tolerance is necessary. The methodology illustrated in this Thesis is applied to a six-component C3-C4 splitter through simulation examples in order to be assessed on the basis of the structural results, according to different control and estimation objectives.
Both control and estimation problems are solved through a unified methodology that consists in the partition of both problems into a structural and an algorithmic part, puts together techniques already employed and new ideas, and permits a systematic design. As regards the control part, the structure is defined as the joint selection of (i) the temperatures to be regulated and (ii) their pairing with the chosen manipulated variables, while the algorithm is the dynamic data processor that processes temperature measurements in order to perform the control task. On the other hand, regarding the estimation part, the structure is defined as the joint selection of (i) a (possibly reduced) estimation model, (ii) the number and locations of temperature measurements, (iii) the states to be innovated through injection of temperature measurements and (iv) the regions in which the temperature can be approximated as the average one, while the algorithm is the dynamic data processor that processes temperature measurements in order to perform the estimation task.
In this Thesis, we propose the temperature gradient with per-component contribution diagram as the mean for the selection of part of the control structure and of the entire estimation structure. Such diagram is the main tool of the used methodology, and consists in plotting temperature gradients at a nominal steady-state together with the contributions due to the components. The diagram is based on the well-known temperature slope criterion, which has extensively been used in control but, as will be shown in this Thesis, provides a simpler and more intuitive criterion for both control and estimation purposes in a multicomponent column, compared to other well-known methods.
As concerns the control algorithm selection, conventional and cascade-type con-trollers with the assistance of first-order observers are developed on the basis of passivity concepts. Such controllers provide a systematic and easy tuning criteria and an anti-windup protection. Their employment is here extended to a multicomponent case.
Geometric Estimator (GE) (with partial innovation) and Extended Kalman Filter (EKF) (with complete or partial innovation) based algorithms are chosen in order to carry out the estimation task. The GE approach, which provides a systematic tuning criterion, is here extended to a multicomponent case
Structural and morphological characterization of hydrozincite and its interaction with organic molecules
Get PDFThe morphological diversity and complexity of naturally occurring forms and patterns have been a motivation for humans to copy and adopt ideas from Nature to achieve functional, aesthetic and social value. Common biogenic materials, such as biominerals, constructed with mineral phases and an organic matrix exhibit delicate structure with unusual optimal chemical-physical, morphological and mechanical properties, and have inspired ideas for the design and synthesis of biomimetic functional materials. The organic templates play an important role in directing assembly of the mineral/organic composites, and in controlling the nucleation and subsequent crystallization. Consequently, knowledge about the molecular interactions at mineral/organic interfaces is essential for understanding the principles of the organic-matrix-mediated biomineralization process.
The microscopic properties of biomineral hydrozincite [Zn₅(CO₃)₂(OH)₆] from Naracauli Creek (SW Sardinia) were investigated by using X-ray Diffraction (XRD), Fourier Transform Infrared (FT-IR), Nuclear Magnetic Resonance (NMR), Scanning Electron Microscopy (SEM), and High-Resolution Transmission Electron Microscopy (HR-TEM) and the organic matrix present has been extracted and characterized by using Fourier Transform Infrared (FT-IR).
In the context of molecular interactions at mineral/organic interfaces we describe an experimental study of the interaction of Bis(2-ethylhexil)phthalate (DEHP) with hydrozincite. This interaction, under controlled laboratory experiments, was investigated by using Fourier Transform Infrared (FT-IR) and Nuclear Magnetic Resonance Spectroscopy (NMR)