Browsing by Author "Fortino, Giancarlo"

Now showing 1 - 12 of 12

<> methodology for the development of autonomic and cognitive internet of things ecosystems
(2018-06-08) Savaglio, Claudio; Crupi, Felice; Fortino, Giancarlo
Advancements on microelectromechanical systems, embedded technologies, and wireless communications have recently enabled the evolution of conven- tional everyday things in enhanced entities, commonly de ned Smart Objects (SOs). Their continuous and widespread di usion, along with an increasing and pervasive connectivity, is enabling unforeseen interactions with conven- tional computing systems, places, animals and humans, thus fading the bound- ary between physical and digital worlds. The Internet of Things (IoT) term just refers to such futuristic scenario, namely a loosely coupled, decentralized and dynamic ecosystem in which bil- lions (even trillions) of self-steering SOs are globally interconnected becoming active participants in business, logistics, information and social processes. In- deed, SOs are able to provide highly pervasive cyberphysical services to both humans and machines thanks to their communication, sensing, actuation, and embedded processing capabilities. Nowadays, the systemic revolution that can be led through the complete realization of the IoT vision is just at its dawn. As matter of facts, whereas new IoT devices and systems have been already developed, they often result in poorly interoperating \Intra-nets of things", mainly due to the heterogeneity featuring IoT building blocks and the lack of standards. Thus, the develop- ment of massive scaled (the total number of \things" is forecasted to reach 20.4 billion in 2020) and actually interoperable IoT systems is a challenging task, featured by several requirements and novel, even unsurveyed, issues. In this context, a multidisciplinary and systematic development approach is necessary, so to involve di erent elds of expertise for coping with the cy- berphysical nature of IoT ecosystem. Henceforth, full- edged IoT methodolo- gies are gaining traction, aiming at systematically supporting all development phases, addressing mentioned issues, and reducing time-to-market, e orts and probability of failure. In such a scenario, this Thesis proposes an application domain-neutral, full- edged agent-based development methodology able to support the main engineering phases of IoT ecosystems. The de nition of such systematic approach resulted in ACOSO-Meth (Agent-based COoperating Smart Objects Methodology), which is the major contribution of this thesis along with other interesting research e orts supporting (i.e., a multi-technology and multi- protocol smartphone-based IoT gateway) and extending (i.e., a full- edged approach to the IoT services modeling according to their opportunistic prop- erties) the main proposal. Finally, to provide validation and performance eval- uation of the proposed ACOSO-Meth approach, four use cases (related to di erent application contexts such as a smart university campus, a smart dig- ital library, a smart city and a smart workshop) have been developed. These research prototypes showed the e ectiveness and e ciency of the proposed approach and improved their respective state-of-the-art. ii
Autonomic computing-based wireless sensor networks
(2013-11-27) Galzarano, Stefano; Fortino, Giancarlo; Liotta, Antonio; Greco, Sergio
Wireless Sensor Networks (WSNs) have grown in popularity in the last years by proving to be a bene cial technology for a wide range of application do- mains, including but not limited to health-care, environment and infrastruc- ture monitoring, smart home automation, industrial control, intelligent agri- culture, and emergency management. However, developing applications on such systems requires many e orts due to the lack of proper software abstractions and the di culties in man- aging resource-constrained embedded environments. Moreover, these appli- cations have to meet a combination of con icting requirements. Achieving accuracy, e ciency, correctness, fault-tolerance, adaptability and reliability on WSN is a major issue because these features have to be provided beyond the design/implementation phase, notably at execution time. This thesis explores the viability and convenience of Autonomic Comput- ing in the context of WSNs by providing a novel paradigm to support the development of autonomic WSN applications as well as speci c self-adaptive protocols at networking levels. In particular, this thesis provides three main contributions. The rst is the design and realization of a novel framework for the development of e cient distributed signal processing applications on heterogeneous WSNs, called SPINE2. It provides a programming abstraction based on the task-oriented paradigm for abstracting away low-level details and has a platform-independent architecture enabling code reusability and portability, application interoperability and platform heterogeneity. The sec- ond contribution is the development of SPINE-* which is an enhancement of SPINE2 by means of an autonomic plane, a way for separating out the provision of self-* techniques from the WSN application logic. Such a separa- tion of concerns leads to an ease of deployment and run-time management of new applications. We nd that this enhancement brings not only considerable functional improvements but also measurable performance bene ts. Third, since we advocate that the agent-oriented paradigm is a well-suited approach in the context of autonomic computing, we propose MAPS, an agent-based programming framework for WSNs. Speci cally designed for supporting Java- iii based sensor platforms, MAPS allows the development of general-purpose mobile multi-agent applications by adopting a multi-plane state machine for- malism for de ning agents' behavior. Finally, the fourth contribution regards the design, analysis, and simulations of a self-adaptive AODV routing protocol enhancement, CG-AODV, and a novel contention-based MAC protocol, QL- MAC. CG-AODV adopts a \node concentration-driven gossiping" approach for limiting the ooding of control packets, whereas QL-MAC, based on a Q-learning approach, aims to nd an e cient radio wake-up/sleep scheduling strategy to reduce energy consumption on the basis of the actual network load of the neighborhood. Simulation results show that CG-AODV outper- forms AODV, whereas QL-MAC provides better performance over standard MAC protocols.
Design Methodologies for FPGA-based Deep Learning Accelerators and Their Characterization
(Università della Calabria, 2023) Sestito, Cristian; Fortino, Giancarlo; Perri, Stefania; Corsonello, Pasquale
Deep Neural Networks (DNNs) are widespread in many applications, including computer vision, speech recognition and robotics, thanks to the ability of such models to extract information by building a hierarchical representation of knowledge. Image processing benefits from the latter behavior by using Convolutional Neural Networks (CNNs), which consist of several Convolutional (CONV) layers to extract features from inputs at different levels of abstraction. However, CNNs usually require billions of computations to reach high accuracy levels. In order to sustain such computational load, proper hardware acceleration is needed. Field Programmable Gate Arrays (FPGAs) have been shown as promising candidates, because they are able to achieve high throughput at limited power dissipation. In addition, FPGAs are flexible architectures to accommodate several CNNs’ workloads. While the hardware acceleration of conventional CNN models has been widely investigated, the interest about more sophisticated tasks is still emerging. The latter includes CNNs based on Dilated Convolutions (DCONVs) and Transposed Convolutions (TCONVs), which deal with filter and image dilations, respectively. Accordingly, higher computational complexity is exhibited by these architectures, thus requiring careful hardware management. This PhD dissertation deals with the FPGA acceleration of CNNs for Image Processing based on DCONVs and TCONVs. Specifically, several designs using both the Very High-Speed Integrated Circuits Hardware Description Language (VHDL) and the High-Level Synthesis (HLS) are presented. Detailed characterization is discussed, based on the evaluation of resources occupation, throughput, power dissipation, as well as the impact of data quantization. Overall, the proposed circuits show noticeable energyefficiency when compared to several state-of-the-art counterparts. For instance, hardware acceleration of run-time reconfigurable CONVs and TCONVs for super-resolution imaging has shown an energy-efficiency of up to 518.5 GOPS/W, by outperforming stateof- the-art competitors by up to 2.3 times.
Design of physically unclonable functions in cmos and emerging technologies for hardware security applications
(Università della Calabria, 2023-02-23) Vatalaro, Massimo; Fortino, Giancarlo; Crupi, Felice
The advent of the IoT scenario heavily pushed the demand of preserving the information down to the chip level due to the increasing demand of interconnected devices. Novel algorithms and hardware architectures are developed every year with the aim of making these systems more and more secure. However, IoT devices operate with constrained area, energy and budget thus making the hardware implementation of these architectures not always feasible. Moreover, these algorithms require truly random key for guarantying a certain security degree. Typically, these secret keys are generated off chip and stored in a non-volatile manner. Unfortunately, this approach requires additional costs and suffers from reverse engineering attacks. Physically unclonable functions (PUFs) are emerging cryptographic primitives which exploit random phenomena, such as random process variations in CMOS manufacturing processes, for generating a unique, repeatable, random, and secure keys in a volatile manner, like a digital fingerprint. PUFs represent a secure and low-cost solution for implementing lightweight cryptographic algorithms. Ideally PUF data should be unique and repeatable even under noisy or different environmental conditions. Unfortunately, guarantying a proper stability is still challenging, especially under PVT variations, thus requiring stability enhancement techniques which overtake the PUF itself in terms of required area and energy. Nowadays, different PUF solutions have been proposed with the aim of achieving ever more stable responses while keeping the area overhead low. This thesis presents a novel class of static monostable PUFs based on a voltage divider between two nominally identical sub-circuits. The fully static behavior along with the use of nominally identical sub-circuits ensure that the correct output is always delivered even when on-chip noise occasionally flips the bit, and that randomness is always guaranteed regardless of the PVT conditions. Measurement results in 180-nm CMOS technology demonstrates the effectiveness of the proposed solution with a native instability (BER) of only 0.61% (0.13%) along with a low sensitivity to both temperature and voltage variations. However, these results were achieved at the cost of more area-hungry design (i.e., 7,222𝐹 ) compared to other relevant works. The proposed solution was also implemented with emerging paper based MoS2 nFETs by exploiting a LUT-based Verilog-A model, calibrated with experimental 𝐼 vs 𝑉 at different 𝑉 curves, whose variability was extracted from different 𝐼 vs 𝑉 curves of 27 devices from the same manufacturing lot. Simulations results demonstrate that these devices can potentially used as building block for next generation electronics targeting hardware security applications. Finally, this thesis also provides an application scenario, in which the proposed PUF solution is employed as TRNG module for implementing a smart tag targeting anti-counterfeiting applications.
Distributed Big Social Data Analysis: Advanced Techniques and Execution Strategies
(Università della Calabria, 2023-05-16) Cantini, Riccardo; Fortino, Giancarlo; Trunfio, Paolo; Marozzo, Fabrizio
A Domain-Specific approach for Programming Wireless Body Sensor Network Systems
(2011-11-23) Gravina, Raffaele; Palopoli, Luigi; Fortino, Giancarlo
The progress of science and medicine during the last years has contributed to signi cantly increase the average life expectancy. The increase of elderly population will have a large impact especially on the health care system. Furthermore, especially in more developed countries, there is an always growing interest in maintaining, and improving the quality of life. Wireless Body Sensor Networks (BSNs) can contribute to improve the quality of health care services. BSNs involve wireless wearable physiological sensors applied to the human body for strictly medical and non medical purposes. They can enhance many human-centered application domains such as e-Health, sport and wellness, and even social applications such as physical/ virtual social interactions. However, there are still open issues that limit their wide di usion in real life; primarily, the programming complexity of these systems, due to lack of high-level software abstractions, and to hardware constraints of wearable devices. In contrast to low-level programming and general-purpose middleware, domain-speci c frameworks are an emerging programming paradigm designed to ful ll the lack of suitable BSN programming support. With this aim, this thesis proposes a novel domain-speci c approach for programming signal-processing intensive BSN applications. The de nition of this approach resulted in a domain-speci c programming framework named SPINE (Signal Processing in Node Environment) which is one important contribution of this thesis, along with other interesting contributions derived from enhancements and variants to the main proposal. Additionally, to provide validation and performance evaluation of the proposed approach, a number of BSN applications (including human activity monitoring, physical energy expenditure estimation, emotional stress detection, and step-counting) have been developed atop SPINE. These research prototypes showed the e ectiveness and e ciency of the proposed approach and improved their respective state-of-the-art. Finally, a Platform-Based Design (PBD) methodology, which is widely adopted for the design of traditional embedded systems, is proposed for the design of BSN systems.
Ensemble of deep learning prediction models for data analytics
(Università della Calabria, 2021-06-21) Zicari, Paolo; Fortino, Giancarlo; Folino, Gianluigi
The abundance of available unstructured or raw text requires the automatic extraction of information for di↵erent tasks. One of the most relevant, Text Classification, extracts this information by assigning informative labels to raw texts from a pre-defined set. Deep Learning (DL) o↵ers challenging solutions to the automatic text classification problem. Despite the great potentialities of DL-based text classifiers, current solutions are exposed to a number of challenging issues that frequently occur in scenarios where text categorization is used in reallife applications. First of all, a large number of labelled data are usually necessary to train a deep model adequately, while labelling texts is timeconsuming, expensive, and very often requires specific knowledge. Moreover, configuring the structure and hyper-parameters of a Deep Neural Network (DNN) architecture is a difficult task, which entails long and careful design and tuning activities to make the DNN perform well. Typical scenarios are characterized by the fact that classes are often imbalanced. These issues entail a high risk of eventually obtaining a DNN-based classifier that overfits the training data and relies on non-general, biased and unreliable classification patterns. On the other hand, the black-box nature of a DNN model does not allow for easy reasoning on which features of a data instance drove the model to its classification decision. The work in this thesis, starting from the general problem of text classification, focuses on some challenging aspects associated with using an ensemble of deep learning methods to classify raw texts. More in detail, this work focuses on the analysis, exploration, study and test of algorithms and learning models to be employed in the proposal of novel techniques of Ensemble Deep Learning (EDL) aimed at performing classification and explanation tasks and on the research of semi-supervised strategies based on pseudo-labelling for improving classifier prediction performances in case of scarcity of labelled data. To this aim, this thesis proposes a complete framework based on the paradigm of ensembles of deep learning algorithms. The proposed framework is designed to furnish a valid instrument for exploring, validating and testing the proposed novel deep ensemble techniques contextualised in reallife applications, covering the entire classification process, including preprocessing, learning model building, explanation of the results, self-training for scarce labelled data, human-in-the-loop validating and model refining. Even though the methods proposed in this work could be used in any field of interest, the problem of extracting information from the raw text was specialised for two specific application contexts: automatic customer support ticket classification and the problem of fake detection. The first application scenario deals with the necessity of the Customer Care Department of most companies to answer their customer requests applied as tickets through several common channels like email, short message texts, social posts, etc. Ticket classification is necessary for automatic answer generation and routing to the specific human operator. Limiting the spread of misinformation, related to the high growth of social media dissemination and sharing of information, has raised the issue of distinguishing true news from fakes, with the challenging problem of processing long texts like news for fake detection. For this reason, the second scenario deals with the critical problem of discerning fake news from the vast amount of information circulating on the Web. In these research areas, the ensemble paradigm has been adopted only recently; thus, discovering the possible advantages when applying this technique is challenging. Experimental tests conducted on real data collected by two Customer Relationship Management (CRM) systems have proven the framework’s effectiveness in di↵erent ticket categorisation tasks and the practical value of their associated explanations. In addition, experiments conducted on two fake news datasets have proven the e↵ectiveness of the proposed semisupervised self-training ensemble-based strategy for improving performances when a few labelled data are available.
Feasibility of a frequency-modulated, wireless, MEMS acceleration evaluator (ALE) for the measurement of lowfrequency and low-amplitude vibrations
(2014-10-28) Sabato, Alessandro; Pagnotta, Leonardo; Oliveti, Giuseppe; Fortino, Giancarlo; Feng, Maria Q.
La necessità di monitorare le vibrazioni sulle strutture e su elementi meccanici assume notevole importanza. Mantenere sotto controllo strutture vetuste, infrastrutture e macchinari ricopre un ruolo strategico nella prevenzioni di situazioni che potrebbero destabilizzare, o finanche distruggere, le strutture stesse. Recenti sviluppi nei sistemi Micro Elettro Meccanici (MEMS), hanno reso gli accelerometri MEMS uno strumento attraente per il monitoraggio dello stato in cui versano diverse tipologie di strutture (SHM). Fino ad ora, la scarsa sensibilità e la bassa risoluzione dei sensori adoperati, specie alle bassissime frequenze, hanno costituito una seria limitazione per il monitoraggio di grandi strutture. Convenzionalmente, i segnali analogici in uscita dagli accelerometri MEMS sono convertiti in segnali digitali prima di essere trasmessi tramite radio-frequenza (RF). La conversione può causare perdita di accuratezza nell’analisi dei segnali caratterizzati da bassa frequenza e piccola ampiezza, che sono d’interesse nel campo dello SHM. Al fine di ovviare a tale limitazione, un nuovo sistema per la valutazione di accelerazioni (ALE) è progettato e realizzato in questo studio. ALE converte il valore di tensione in uscita dall’accelerometro MEMS in un segnale modulato in frequenza (FM) prima della radiotrasmissione. Per fare ciò, è adoperato, al posto del tradizionale Convertitore Analogico Digitale (ADC), un convertitore Voltaggio – Frequenza (V/F). Il prototipo realizzato in questo studio è formato da due schede: una trasmittente e una ricevente. La prima unità è equipaggiata con il suddetto accelerometro MEMS, un convertitore V/F e un’antenna trasmittente; la seconda, invece, per demodulare il segnale ed inviarlo ad una scheda di acquisizione dati (DAQ) collegata ad un computer, utilizza un’antenna ricevente e un convertitore Frequenza – Voltaggio (F/V). L’efficacia del prototipo realizzato, nel misurare sollecitazioni dinamiche caratterizzate da bassissime frequenze e piccole ampiezze, è valutata per dimostrare la possibilità di utilizzo di ALE nelle analisi SHM. Per fare ciò, sono state effettuate prove di laboratorio ed esperimenti su differenti strutture reali. Le prime consistono in calibrazioni statiche, studi sugli effetti che la carica residua delle batterie può avere sul comportamento delle schede, prove per valutare la distanza massima alla quale il sistema RF può trasmettere, e una serie di misure comparative con accelerometri cablati. Le seconde, invece, consistono in una serie di misurazioni effettuate su diverse tipologie di sistemi dinamici: una condotta per il trasporto di carburante (applicazioni industriali), una guglia in pietra di un edificio storico (applicazioni sismiche) e un ponte pedonale (monitoraggio di infrastrutture civili). In questo studio sono discussi i risultati raggiunti e fornite indicazioni utili per possibili futuri sviluppi del prototipo.
Feature Selection in Classification by means of Optimization and Multi-Objective Optimization
(Università della Calabria, 2023-05-10) Pirouz, Behzad; Fortino, Giancarlo; Gaudioso, Manlio
The thesis is in the area of mathematical optimization with application to Machine Learning. The focus is on Feature Selection (FS) in the framework of binary classification via Support Vector Machine paradigm. We concentrate on the use of sparse optimization techniques, which are widely considered as the election tool for tackling FS. We study the problem both in terms of single and multi-objective optimization. We propose first a novel Mixed-Integer Nonlinear Programming (MINLP) model for sparse optimization based on the polyhedral k-norm. We introduce a new way to take into account the k-norm for sparse optimization by setting a model based on fractional programming (FP). Then we address the continuous relaxation of the problem, which is reformulated via a DC (Difference of Convex) decomposition. On the other hand, designing supervised learning systems, in general, is a multi-objective problem. It requires finding appropriate trade-offs between several objectives, for example, between the number of misclassified training data (minimizing the squared error) and the number of nonzero elements separating the hyperplane (minimizing the number of nonzero elements). When we deal with multi-objective optimization problems, the optimization problem has yet to have a single solution that represents the best solution for all objectives simultaneously. Consequently, there is not a single solution but a set of solutions, known as the Pareto-optimal solutions. We overview the SVM models and the related Feature Selection in terms of multi-objective optimization. Our multi-objective approach considers two simultaneous objectives: minimizing the squared error and minimizing the number of nonzero elements of the normal vector of the separator hyperplane. In this thesis, we propose a multi-objective model for sparse optimization. Our primary purpose is to demonstrate the advantages of considering SVM models as multi-objective optimization problems. In multi-objective cases, we can obtain a set of Pareto optimal solutions instead of one in single-objective cases. Therefore, our main contribution in this thesis is of two levels: first, we propose a new model for sparse optimization based on the polyhedral k-norm for SVM classification, and second, use multi-objective optimization to consider this new model. The results of several numerical experiments on some classification datasets are reported. We used all the datasets for single-objective and multi-objective models.
High-level frameworks for the development of wireless sensor network applications
(2011-11-23) Guerrieri, Antonio; Palopoli, Luigi; Fortino, Giancarlo
Wireless Sensor Networks (WSNs) are emerging as powerful platforms for distributed embedded computing supporting a variety of high-impact appli- cations. A WSN is a group of small devices (nodes) capable to sample the real world through sensors, actuate commands through actuators, elaborate data on the node, and send messages to other nodes through radio communi- cation. However, programming WSN applications is a complex task that re- quires suitable paradigms and technologies capable of supporting the speci c characteristics of such networks which uniquely integrate distributed sensing, computation and communication. This thesis aims at providing new paradigms to support the development of WSN applications through both a domain-speci c and a general-purpose approach. In particular, this thesis provides three main contributions. The rst is related to the analysis, design and realization of a domain-speci c frame- work for heterogeneous WSNs for exible and e cient distributed sensing and actuation in buildings called Building Management Framework (BMF). BMF provides fast WSN recon guration, in-node processing algorithms, multi-hop networks, and multi-platform support, a programming abstraction to dynami- cally catch the morphology of buildings, actuators support, and an extensible human computer interface. The second contribution refers to the analysis, design and realization of a general-purpose mobile agent system for WSN, namely MAPS (Multi Agent Platform for SunSPOT). MAPS allows an e ec- tive Java-based development of agents and agent-based applications for WSNs by integrating agent oriented, event-driven and state-based programming pa- radigms. Finally, the third contribution regards the analysis, design and re- alization of a domain-speci c framework for rapid prototyping of platform independent Wireless Body Sensor Network (WBSN) applications, namely SPINE2 (signal processing in-node environment version 2). SPINE2 aims at supporting the development of WSN applications raising the level of the used programming abstractions by providing a task-oriented programming model.
Inverse source and scattering solutions for microwave imaging applications
(Università della Calabria, 2023-02-02) Lopez, Giuseppe; Fortino, Giancarlo; Costanzo, Sandra
A logical and ontological framework for metadata extraction and modelling from heterogeneous document sources
(Università della Calabria, 2023-06-24) Cuconato, Simone; Fortino, Giancarlo; Folino, Antonietta; Cardillo, Elena