Statistical methods for biomedical signal analysis and processing

Statistical modelling and statistical learning theory are two powerful analytical frameworks for analyzing signals and developing efficient processing and classification algorithms. In this thesis, these frameworks are applied for modelling and processing biomedical signals in two different contexts: ultrasound medical imaging systems and primate neural activity analysis and modelling. In the context of ultrasound medical imaging, two main applications are explored: deconvolution of signals measured from a ultrasonic transducer and automatic image segmentation and classification of prostate ultrasound scans. In the former application a stochastic model of the radio frequency signal measured from a ultrasonic transducer is derived. This model is then employed for developing in a statistical framework a regularized deconvolution procedure, for enhancing signal resolution. In the latter application, different statistical models are used to characterize images of prostate tissues, extracting different features. These features are then uses to segment the images in region of interests by means of an automatic procedure based on a statistical model of the extracted features. Finally, machine learning techniques are used for automatic classification of the different region of interests. In the context of neural activity signals, an example of bio-inspired dynamical network was developed to help in studies of motor-related processes in the brain of primate monkeys. The presented model aims to mimic the abstract functionality of a cell population in 7a parietal region of primate monkeys, during the execution of learned behavioural tasks.

Fault detection in rotating machines by vibration signal processing techniques

Machines with moving parts give rise to vibrations and consequently noise. The setting up and the status of each machine yield to a peculiar vibration signature. Therefore, a change in the vibration signature, due to a change in the machine state, can be used to detect incipient defects before they become critical. This is the goal of condition monitoring, in which the informations obtained from a machine signature are used in order to detect faults at an early stage. There are a large number of signal processing techniques that can be used in order to extract interesting information from a measured vibration signal. This study seeks to detect rotating machine defects using a range of techniques including synchronous time averaging, Hilbert transform-based demodulation, continuous wavelet transform, Wigner-Ville distribution and spectral correlation density function. The detection and the diagnostic capability of these techniques are discussed and compared on the basis of experimental results concerning gear tooth faults, i.e. fatigue crack at the tooth root and tooth spalls of different sizes, as well as assembly faults in diesel engine. Moreover, the sensitivity to fault severity is assessed by the application of these signal processing techniques to gear tooth faults of different sizes.

Adaptive multiscale biological signal processing

Biological processes are very complex mechanisms, most of them being accompanied by or manifested as signals that reflect their essential characteristics and qualities. The development of diagnostic techniques based on signal and image acquisition from the human body is commonly retained as one of the propelling factors in the advancements in medicine and biosciences recorded in the recent past. It is a fact that the instruments used for biological signal and image recording, like any other acquisition system, are affected by non-idealities which, by different degrees, negatively impact on the accuracy of the recording. This work discusses how it is possible to attenuate, and ideally to remove, these effects, with a particular attention toward ultrasound imaging and extracellular recordings. Original algorithms developed during the Ph.D. research activity will be examined and compared to ones in literature tackling the same problems; results will be drawn on the base of comparative tests on both synthetic and in-vivo acquisitions, evaluating standard metrics in the respective field of application. All the developed algorithms share an adaptive approach to signal analysis, meaning that their behavior is not dependent only on designer choices, but driven by input signal characteristics too. Performance comparisons following the state of the art concerning image quality assessment, contrast gain estimation and resolution gain quantification as well as visual inspection highlighted very good results featured by the proposed ultrasound image deconvolution and restoring algorithms: axial resolution up to 5 times better than algorithms in literature are possible. Concerning extracellular recordings, the results of the proposed denoising technique compared to other signal processing algorithms pointed out an improvement of the state of the art of almost 4 dB.

Heterogeneous multicore systems for signal processing

This thesis explores the capabilities of heterogeneous multi-core systems, based on multiple Graphics Processing Units (GPUs) in a standard desktop framework. Multi-GPU accelerated desk side computers are an appealing alternative to other high performance computing (HPC) systems: being composed of commodity hardware components fabricated in large quantities, their price-performance ratio is unparalleled in the world of high performance computing. Essentially bringing “supercomputing to the masses”, this opens up new possibilities for application fields where investing in HPC resources had been considered unfeasible before. One of these is the field of bioelectrical imaging, a class of medical imaging technologies that occupy a low-cost niche next to million-dollar systems like functional Magnetic Resonance Imaging (fMRI). In the scope of this work, several computational challenges encountered in bioelectrical imaging are tackled with this new kind of computing resource, striving to help these methods approach their true potential. Specifically, the following main contributions were made: Firstly, a novel dual-GPU implementation of parallel triangular matrix inversion (TMI) is presented, addressing an crucial kernel in computation of multi-mesh head models of encephalographic (EEG) source localization. This includes not only a highly efficient implementation of the routine itself achieving excellent speedups versus an optimized CPU implementation, but also a novel GPU-friendly compressed storage scheme for triangular matrices. Secondly, a scalable multi-GPU solver for non-hermitian linear systems was implemented. It is integrated into a simulation environment for electrical impedance tomography (EIT) that requires frequent solution of complex systems with millions of unknowns, a task that this solution can perform within seconds. In terms of computational throughput, it outperforms not only an highly optimized multi-CPU reference, but related GPU-based work as well. Finally, a GPU-accelerated graphical EEG real-time source localization software was implemented. Thanks to acceleration, it can meet real-time requirements in unpreceeded anatomical detail running more complex localization algorithms. Additionally, a novel implementation to extract anatomical priors from static Magnetic Resonance (MR) scansions has been included.

Ultra-low power WSNs: distributed signal processing and dynamic resource management

This thesis presents several data processing and compression techniques capable of addressing the strict requirements of wireless sensor networks. After introducing a general overview of sensor networks, the energy problem is introduced, dividing the different energy reduction approaches according to the different subsystem they try to optimize. To manage the complexity brought by these techniques, a quick overview of the most common middlewares for WSNs is given, describing in detail SPINE2, a framework for data processing in the node environment. The focus is then shifted on the in-network aggregation techniques, used to reduce data sent by the network nodes trying to prolong the network lifetime as long as possible. Among the several techniques, the most promising approach is the Compressive Sensing (CS). To investigate this technique, a practical implementation of the algorithm is compared against a simpler aggregation scheme, deriving a mixed algorithm able to successfully reduce the power consumption. The analysis moves from compression implemented on single nodes to CS for signal ensembles, trying to exploit the correlations among sensors and nodes to improve compression and reconstruction quality. The two main techniques for signal ensembles, Distributed CS (DCS) and Kronecker CS (KCS), are introduced and compared against a common set of data gathered by real deployments. The best trade-off between reconstruction quality and power consumption is then investigated. The usage of CS is also addressed when the signal of interest is sampled at a Sub-Nyquist rate, evaluating the reconstruction performance. Finally the group sparsity CS (GS-CS) is compared to another well-known technique for reconstruction of signals from an highly sub-sampled version. These two frameworks are compared again against a real data-set and an insightful analysis of the trade-off between reconstruction quality and lifetime is given.

Efficient ultrasonic signal processing techniques for aided medical diagnostics

Ultrasound imaging is widely used in medical diagnostics as it is the fastest, least invasive, and least expensive imaging modality. However, ultrasound images are intrinsically difficult to be interpreted. In this scenario, Computer Aided Detection (CAD) systems can be used to support physicians during diagnosis providing them a second opinion. This thesis discusses efficient ultrasound processing techniques for computer aided medical diagnostics, focusing on two major topics: (i) Ultrasound Tissue Characterization (UTC), aimed at characterizing and differentiating between healthy and diseased tissue; (ii) Ultrasound Image Segmentation (UIS), aimed at detecting the boundaries of anatomical structures to automatically measure organ dimensions and compute clinically relevant functional indices. Research on UTC produced a CAD tool for Prostate Cancer detection to improve the biopsy protocol. In particular, this thesis contributes with: (i) the development of a robust classification system; (ii) the exploitation of parallel computing on GPU for real-time performance; (iii) the introduction of both an innovative Semi-Supervised Learning algorithm and a novel supervised/semi-supervised learning scheme for CAD system training that improve system performance reducing data collection effort and avoiding collected data wasting. The tool provides physicians a risk map highlighting suspect tissue areas, allowing them to perform a lesion-directed biopsy. Clinical validation demonstrated the system validity as a diagnostic support tool and its effectiveness at reducing the number of biopsy cores requested for an accurate diagnosis. For UIS the research developed a heart disease diagnostic tool based on Real-Time 3D Echocardiography. Thesis contributions to this application are: (i) the development of an automated GPU based level-set segmentation framework for 3D images; (ii) the application of this framework to the myocardium segmentation. Experimental results showed the high efficiency and flexibility of the proposed framework. Its effectiveness as a tool for quantitative analysis of 3D cardiac morphology and function was demonstrated through clinical validation.

Diagnosis and Fault detection in Electrical Machines and Drives based on Advanced Signal Processing Techniques

In the present thesis, a new methodology of diagnosis based on advanced use of time-frequency technique analysis is presented. More precisely, a new fault index that allows tracking individual fault components in a single frequency band is defined. More in detail, a frequency sliding is applied to the signals being analyzed (currents, voltages, vibration signals), so that each single fault frequency component is shifted into a prefixed single frequency band. Then, the discrete Wavelet Transform is applied to the resulting signal to extract the fault signature in the frequency band that has been chosen. Once the state of the machine has been qualitatively diagnosed, a quantitative evaluation of the fault degree is necessary. For this purpose, a fault index based on the energy calculation of approximation and/or detail signals resulting from wavelet decomposition has been introduced to quantify the fault extend. The main advantages of the developed new method over existing Diagnosis techniques are the following: - Capability of monitoring the fault evolution continuously over time under any transient operating condition; - Speed/slip measurement or estimation is not required; - Higher accuracy in filtering frequency components around the fundamental in case of rotor faults; - Reduction in the likelihood of false indications by avoiding confusion with other fault harmonics (the contribution of the most relevant fault frequency components under speed-varying conditions are clamped in a single frequency band); - Low memory requirement due to low sampling frequency; - Reduction in the latency of time processing (no requirement of repeated sampling operation).

Embedded systems and advanced signal processing for Acousto- Ultrasonic Inspections

Non Destructive Testing (NDT) and Structural Health Monitoring (SHM) are becoming essential in many application contexts, e.g. civil, industrial, aerospace etc., to reduce structures maintenance costs and improve safety. Conventional inspection methods typically exploit bulky and expensive instruments and rely on highly demanding signal processing techniques. The pressing need to overcome these limitations is the common thread that guided the work presented in this Thesis. In the first part, a scalable, low-cost and multi-sensors smart sensor network is introduced. The capability of this technology to carry out accurate modal analysis on structures undergoing flexural vibrations has been validated by means of two experimental campaigns. Then, the suitability of low-cost piezoelectric disks in modal analysis has been demonstrated. To enable the use of this kind of sensing technology in such non conventional applications, ad hoc data merging algorithms have been developed. In the second part, instead, imaging algorithms for Lamb waves inspection (namely DMAS and DS-DMAS) have been implemented and validated. Results show that DMAS outperforms the canonical Delay and Sum (DAS) approach in terms of image resolution and contrast. Similarly, DS-DMAS can achieve better results than both DMAS and DAS by suppressing artefacts and noise. To exploit the full potential of these procedures, accurate group velocity estimations are required. Thus, novel wavefield analysis tools that can address the estimation of the dispersion curves from SLDV acquisitions have been investigated. An image segmentation technique (called DRLSE) was exploited in the k-space to draw out the wavenumber profile. The DRLSE method was compared with compressive sensing methods to extract the group and phase velocity information. The validation, performed on three different carbon fibre plates, showed that the proposed solutions can accurately determine the wavenumber and velocities in polar coordinates at multiple excitation frequencies.

Assessing brain connectivity through electroencephalographic signal processing and modeling analysis

Brain functioning relies on the interaction of several neural populations connected through complex connectivity networks, enabling the transmission and integration of information. Recent advances in neuroimaging techniques, such as electroencephalography (EEG), have deepened our understanding of the reciprocal roles played by brain regions during cognitive processes. The underlying idea of this PhD research is that EEG-related functional connectivity (FC) changes in the brain may incorporate important neuromarkers of behavior and cognition, as well as brain disorders, even at subclinical levels. However, a complete understanding of the reliability of the wide range of existing connectivity estimation techniques is still lacking. The first part of this work addresses this limitation by employing Neural Mass Models (NMMs), which simulate EEG activity and offer a unique tool to study interconnected networks of brain regions in controlled conditions. NMMs were employed to test FC estimators like Transfer Entropy and Granger Causality in linear and nonlinear conditions. Results revealed that connectivity estimates reflect information transmission between brain regions, a quantity that can be significantly different from the connectivity strength, and that Granger causality outperforms the other estimators. A second objective of this thesis was to assess brain connectivity and network changes on EEG data reconstructed at the cortical level. Functional brain connectivity has been estimated through Granger Causality, in both temporal and spectral domains, with the following goals: a) detect task-dependent functional connectivity network changes, focusing on internal-external attention competition and fear conditioning and reversal; b) identify resting-state network alterations in a subclinical population with high autistic traits. Connectivity-based neuromarkers, compared to the canonical EEG analysis, can provide deeper insights into brain mechanisms and may drive future diagnostic methods and therapeutic interventions. However, further methodological studies are required to fully understand the accuracy and information captured by FC estimates, especially concerning nonlinear phenomena.

Data Mining in Clinical Practice for the Quantification of Motor Impairment in Parkinson's Disease

Advances in biomedical signal acquisition systems for motion analysis have led to lowcost and ubiquitous wearable sensors which can be used to record movement data in different settings. This implies the potential availability of large amounts of quantitative data. It is then crucial to identify and to extract the information of clinical relevance from the large amount of available data. This quantitative and objective information can be an important aid for clinical decision making. Data mining is the process of discovering such information in databases through data processing, selection of informative data, and identification of relevant patterns. The databases considered in this thesis store motion data from wearable sensors (specifically accelerometers) and clinical information (clinical data, scores, tests). The main goal of this thesis is to develop data mining tools which can provide quantitative information to the clinician in the field of movement disorders. This thesis will focus on motor impairment in Parkinson's disease (PD). Different databases related to Parkinson subjects in different stages of the disease were considered for this thesis. Each database is characterized by the data recorded during a specific motor task performed by different groups of subjects. The data mining techniques that were used in this thesis are feature selection (a technique which was used to find relevant information and to discard useless or redundant data), classification, clustering, and regression. The aims were to identify high risk subjects for PD, characterize the differences between early PD subjects and healthy ones, characterize PD subtypes and automatically assess the severity of symptoms in the home setting.

Studio neuroradiologico (morfologico e funzionale) nei pazienti con Epilessia Frontale Notturna

L'epilessia frontale notturna (EFN) è caratterizzata da crisi motorie che insorgono durante il sonno. Scopo del progetto è studiare le cause fisiopatologiche e morfo-funzionali che sottendono ai fenomeni motori nei pazienti con EFN e identificare alterazioni strutturali e/o metaboliche mediante tecniche avanzate di Risonanza Magnetica (RM). Abbiamo raccolto una casistica di pazienti con EFN afferenti al Centro Epilessia e dei Disturbi del Sonno del Dipartimento di Scienze Neurologiche, Università di Bologna. Ad ogni paziente è stato associato un controllo sano di età (± 5 anni) e sesso corrispondente. Tutti sono stati studiati mediante tecniche avanzate di RM comprendenti Spettroscopia del protone (1H-MRS), Tensore di diffusione ed imaging 3D ad alta risoluzione per analisi morfometriche. In particolare, la 1H-MRS è stata effettuata su due volumi di interesse localizzati nei talami e nel giro del cingolo anteriore. Sono stati inclusi nell’analisi finale 19 pazienti (7 M), età media 34 anni (range 19-50) e 14 controlli (6 M) età media 30 anni (range 19-40). A livello del cingolo anteriore il rapporto della concentrazione di N-Acetil-Aspartato rispetto alla Creatina (NAA/Cr) è risultato significativamente ridotto nei pazienti rispetto ai controlli (p=0,021). Relativamente all’analisi di correlazione, l'analisi tramite modelli di regressione multipla ha evidenziato che il rapporto NAA/Cr nel cingolo anteriore nei pazienti correlava con la frequenza delle crisi (p=0,048), essendo minore nei pazienti con crisi plurisettimanali/plurigiornaliere. Per interpretare il dato ottenuto è possibile solo fare delle ipotesi. L’NAA è un marker di integrità, densità e funzionalità neuronale. E’ possibile che alla base della EFN ci siano alterazioni metaboliche tessutali in precise strutture come il giro del cingolo anteriore. Questo apre nuove possibilità sull’utilizzo di strumenti di indagine basati sull’analisi di biosegnali, per caratterizzare aree coinvolte nella genesi della EFN ancora largamente sconosciute e chiarire ulteriormente l’eziologia di questo tipo di epilessia.

Sparse Signal Representation of Ultrasonic Signals for Structural Health Monitoring Applications

Assessment of the integrity of structural components is of great importance for aerospace systems, land and marine transportation, civil infrastructures and other biological and mechanical applications. Guided waves (GWs) based inspections are an attractive mean for structural health monitoring. In this thesis, the study and development of techniques for GW ultrasound signal analysis and compression in the context of non-destructive testing of structures will be presented. In guided wave inspections, it is necessary to address the problem of the dispersion compensation. A signal processing approach based on frequency warping was adopted. Such operator maps the frequencies axis through a function derived by the group velocity of the test material and it is used to remove the dependence on the travelled distance from the acquired signals. Such processing strategy was fruitfully applied for impact location and damage localization tasks in composite and aluminum panels. It has been shown that, basing on this processing tool, low power embedded system for GW structural monitoring can be implemented. Finally, a new procedure based on Compressive Sensing has been developed and applied for data reduction. Such procedure has also a beneficial effect in enhancing the accuracy of structural defects localization. This algorithm uses the convolutive model of the propagation of ultrasonic guided waves which takes advantage of a sparse signal representation in the warped frequency domain. The recovery from the compressed samples is based on an alternating minimization procedure which achieves both an accurate reconstruction of the ultrasonic signal and a precise estimation of waves time of flight. Such information is used to feed hyperbolic or elliptic localization procedures, for accurate impact or damage localization.

TCAD approaches to multidimensional simulation of advanced semiconductor devices

Technology scaling increasingly emphasizes complexity and non-ideality of the electrical behavior of semiconductor devices and boosts interest on alternatives to the conventional planar MOSFET architecture. TCAD simulation tools are fundamental to the analysis and development of new technology generations. However, the increasing device complexity is reflected in an augmented dimensionality of the problems to be solved. The trade-off between accuracy and computational cost of the simulation is especially influenced by domain discretization: mesh generation is therefore one of the most critical steps and automatic approaches are sought. Moreover, the problem size is further increased by process variations, calling for a statistical representation of the single device through an ensemble of microscopically different instances. The aim of this thesis is to present multi-disciplinary approaches to handle this increasing problem dimensionality in a numerical simulation perspective. The topic of mesh generation is tackled by presenting a new Wavelet-based Adaptive Method (WAM) for the automatic refinement of 2D and 3D domain discretizations. Multiresolution techniques and efficient signal processing algorithms are exploited to increase grid resolution in the domain regions where relevant physical phenomena take place. Moreover, the grid is dynamically adapted to follow solution changes produced by bias variations and quality criteria are imposed on the produced meshes. The further dimensionality increase due to variability in extremely scaled devices is considered with reference to two increasingly critical phenomena, namely line-edge roughness (LER) and random dopant fluctuations (RD). The impact of such phenomena on FinFET devices, which represent a promising alternative to planar CMOS technology, is estimated through 2D and 3D TCAD simulations and statistical tools, taking into account matching performance of single devices as well as basic circuit blocks such as SRAMs. Several process options are compared, including resist- and spacer-defined fin patterning as well as different doping profile definitions. Combining statistical simulations with experimental data, potentialities and shortcomings of the FinFET architecture are analyzed and useful design guidelines are provided, which boost feasibility of this technology for mainstream applications in sub-45 nm generation integrated circuits.

Vibration as a method for muscular stimulation

Human reactions to vibration have been extensively investigated in the past. Vibration, as well as whole-body vibration (WBV), has been commonly considered as an occupational hazard for its detrimental effects on human condition and comfort. Although long term exposure to vibrations may produce undesirable side-effects, a great part of the literature is dedicated to the positive effects of WBV when used as method for muscular stimulation and as an exercise intervention. Whole body vibration training (WBVT) aims to mechanically activate muscles by eliciting neuromuscular activity (muscle reflexes) via the use of vibrations delivered to the whole body. The most mentioned mechanism to explain the neuromuscular outcomes of vibration is the elicited neuromuscular activation. Local tendon vibrations induce activity of the muscle spindle Ia fibers, mediated by monosynaptic and polysynaptic pathways: a reflex muscle contraction known as the Tonic Vibration Reflex (TVR) arises in response to such vibratory stimulus. In WBVT mechanical vibrations, in a range from 10 to 80 Hz and peak to peak displacements from 1 to 10 mm, are usually transmitted to the patient body by the use of oscillating platforms. Vibrations are then transferred from the platform to a specific muscle group through the subject body. To customize WBV treatments, surface electromyography (SEMG) signals are often used to reveal the best stimulation frequency for each subject. Use of SEMG concise parameters, such as root mean square values of the recordings, is also a common practice; frequently a preliminary session can take place in order to discover the more appropriate stimulation frequency. Soft tissues act as wobbling masses vibrating in a damped manner in response to mechanical excitation; Muscle Tuning hypothesis suggest that neuromuscular system works to damp the soft tissue oscillation that occurs in response to vibrations; muscles alters their activity to dampen the vibrations, preventing any resonance phenomenon. Muscle response to vibration is however a complex phenomenon as it depends on different parameters, like muscle-tension, muscle or segment-stiffness, amplitude and frequency of the mechanical vibration. Additionally, while in the TVR study the applied vibratory stimulus and the muscle conditions are completely characterised (a known vibration source is applied directly to a stretched/shortened muscle or tendon), in WBV study only the stimulus applied to a distal part of the body is known. Moreover, mechanical response changes in relation to the posture. The transmissibility of vibratory stimulus along the body segment strongly depends on the position held by the subject. The aim of this work was the investigation on the effects that the use of vibrations, in particular the effects of whole body vibrations, may have on muscular activity. A new approach to discover the more appropriate stimulus frequency, by the use of accelerometers, was also explored. Different subjects, not affected by any known neurological or musculoskeletal disorders, were voluntarily involved in the study and gave their informed, written consent to participate. The device used to deliver vibration to the subjects was a vibrating platform. Vibrations impressed by the platform were exclusively vertical; platform displacement was sinusoidal with an intensity (peak-to-peak displacement) set to 1.2 mm and with a frequency ranging from 10 to 80 Hz. All the subjects familiarized with the device and the proper positioning. Two different posture were explored in this study: position 1 - hack squat; position 2 - subject standing on toes with heels raised. SEMG signals from the Rectus Femoris (RF), Vastus Lateralis (VL) and Vastus medialis (VM) were recorded. SEMG signals were amplified using a multi-channel, isolated biomedical signal amplifier The gain was set to 1000 V/V and a band pass filter (-3dB frequency 10 - 500 Hz) was applied; no notch filters were used to suppress line interference. Tiny and lightweight (less than 10 g) three-axial MEMS accelerometers (Freescale semiconductors) were used to measure accelerations of onto patient’s skin, at EMG electrodes level. Accelerations signals provided information related to individuals’ RF, Biceps Femoris (BF) and Gastrocnemius Lateralis (GL) muscle belly oscillation; they were pre-processed in order to exclude influence of gravity. As demonstrated by our results, vibrations generate peculiar, not negligible motion artifact on skin electrodes. Artifact amplitude is generally unpredictable; it appeared in all the quadriceps muscles analysed, but in different amounts. Artifact harmonics extend throughout the EMG spectrum, making classic high-pass filters ineffective; however, their contribution was easy to filter out from the raw EMG signal with a series of sharp notch filters centred at the vibration frequency and its superior harmonics (1.5 Hz wide). However, use of these simple filters prevents the revelation of EMG power potential variation in the mentioned filtered bands. Moreover our experience suggests that the possibility of reducing motion artefact, by using particular electrodes and by accurately preparing the subject’s skin, is not easily viable; even though some small improvements were obtained, it was not possible to substantially decrease the artifact. Anyway, getting rid of those artifacts lead to some true EMG signal loss. Nevertheless, our preliminary results suggest that the use of notch filters at vibration frequency and its harmonics is suitable for motion artifacts filtering. In RF SEMG recordings during vibratory stimulation only a little EMG power increment should be contained in the mentioned filtered bands due to synchronous electromyographic activity of the muscle. Moreover, it is better to remove the artifact that, in our experience, was found to be more than 40% of the total signal power. In summary, many variables have to be taken into account: in addition to amplitude, frequency and duration of vibration treatment, other fundamental variables were found to be subject anatomy, individual physiological condition and subject’s positioning on the platform. Studies on WBV treatments that include surface EMG analysis to asses muscular activity during vibratory stimulation should take into account the presence of motion artifacts. Appropriate filtering of artifacts, to reveal the actual effect on muscle contraction elicited by vibration stimulus, is mandatory. However as a result of our preliminary study, a simple multi-band notch filtering may help to reduce randomness of the results. Muscle tuning hypothesis seemed to be confirmed. Our results suggested that the effects of WBV are linked to the actual muscle motion (displacement). The greater was the muscle belly displacement the higher was found the muscle activity. The maximum muscle activity has been found in correspondence with the local mechanical resonance, suggesting a more effective stimulation at the specific system resonance frequency. Holding the hypothesis that muscle activation is proportional to muscle displacement, treatment optimization could be obtained by simply monitoring local acceleration (resonance). However, our study revealed some short term effects of vibratory stimulus; prolonged studies should be assembled in order to consider the long term effectiveness of these results. Since local stimulus depends on the kinematic chain involved, WBV muscle stimulation has to take into account the transmissibility of the stimulus along the body segment in order to ensure that vibratory stimulation effectively reaches the target muscle. Combination of local resonance and muscle response should also be further investigated to prevent hazards to individuals undergoing WBV treatments.

Tecniche di ottimizzazione del software per sistemi su singolo chip per applicazioni di Nomadic Computing

I moderni sistemi embedded sono equipaggiati con risorse hardware che consentono l’esecuzione di applicazioni molto complesse come il decoding audio e video. La progettazione di simili sistemi deve soddisfare due esigenze opposte. Da un lato è necessario fornire un elevato potenziale computazionale, dall’altro bisogna rispettare dei vincoli stringenti riguardo il consumo di energia. Uno dei trend più diffusi per rispondere a queste esigenze opposte è quello di integrare su uno stesso chip un numero elevato di processori caratterizzati da un design semplificato e da bassi consumi. Tuttavia, per sfruttare effettivamente il potenziale computazionale offerto da una batteria di processoriè necessario rivisitare pesantemente le metodologie di sviluppo delle applicazioni. Con l’avvento dei sistemi multi-processore su singolo chip (MPSoC) il parallel programming si è diffuso largamente anche in ambito embedded. Tuttavia, i progressi nel campo della programmazione parallela non hanno mantenuto il passo con la capacità di integrare hardware parallelo su un singolo chip. Oltre all’introduzione di multipli processori, la necessità di ridurre i consumi degli MPSoC comporta altre soluzioni architetturali che hanno l’effetto diretto di complicare lo sviluppo delle applicazioni. Il design del sottosistema di memoria, in particolare, è un problema critico. Integrare sul chip dei banchi di memoria consente dei tempi d’accesso molto brevi e dei consumi molto contenuti. Sfortunatamente, la quantità di memoria on-chip che può essere integrata in un MPSoC è molto limitata. Per questo motivo è necessario aggiungere dei banchi di memoria off-chip, che hanno una capacità molto maggiore, come maggiori sono i consumi e i tempi d’accesso. La maggior parte degli MPSoC attualmente in commercio destina una parte del budget di area all’implementazione di memorie cache e/o scratchpad. Le scratchpad (SPM) sono spesso preferite alle cache nei sistemi MPSoC embedded, per motivi di maggiore predicibilità, minore occupazione d’area e – soprattutto – minori consumi. Per contro, mentre l’uso delle cache è completamente trasparente al programmatore, le SPM devono essere esplicitamente gestite dall’applicazione. Esporre l’organizzazione della gerarchia di memoria ll’applicazione consente di sfruttarne in maniera efficiente i vantaggi (ridotti tempi d’accesso e consumi). Per contro, per ottenere questi benefici è necessario scrivere le applicazioni in maniera tale che i dati vengano partizionati e allocati sulle varie memorie in maniera opportuna. L’onere di questo compito complesso ricade ovviamente sul programmatore. Questo scenario descrive bene l’esigenza di modelli di programmazione e strumenti di supporto che semplifichino lo sviluppo di applicazioni parallele. In questa tesi viene presentato un framework per lo sviluppo di software per MPSoC embedded basato su OpenMP. OpenMP è uno standard di fatto per la programmazione di multiprocessori con memoria shared, caratterizzato da un semplice approccio alla parallelizzazione tramite annotazioni (direttive per il compilatore). La sua interfaccia di programmazione consente di esprimere in maniera naturale e molto efficiente il parallelismo a livello di loop, molto diffuso tra le applicazioni embedded di tipo signal processing e multimedia. OpenMP costituisce un ottimo punto di partenza per la definizione di un modello di programmazione per MPSoC, soprattutto per la sua semplicità d’uso. D’altra parte, per sfruttare in maniera efficiente il potenziale computazionale di un MPSoC è necessario rivisitare profondamente l’implementazione del supporto OpenMP sia nel compilatore che nell’ambiente di supporto a runtime. Tutti i costrutti per gestire il parallelismo, la suddivisione del lavoro e la sincronizzazione inter-processore comportano un costo in termini di overhead che deve essere minimizzato per non comprometterre i vantaggi della parallelizzazione. Questo può essere ottenuto soltanto tramite una accurata analisi delle caratteristiche hardware e l’individuazione dei potenziali colli di bottiglia nell’architettura. Una implementazione del task management, della sincronizzazione a barriera e della condivisione dei dati che sfrutti efficientemente le risorse hardware consente di ottenere elevate performance e scalabilità. La condivisione dei dati, nel modello OpenMP, merita particolare attenzione. In un modello a memoria condivisa le strutture dati (array, matrici) accedute dal programma sono fisicamente allocate su una unica risorsa di memoria raggiungibile da tutti i processori. Al crescere del numero di processori in un sistema, l’accesso concorrente ad una singola risorsa di memoria costituisce un evidente collo di bottiglia. Per alleviare la pressione sulle memorie e sul sistema di connessione vengono da noi studiate e proposte delle tecniche di partizionamento delle strutture dati. Queste tecniche richiedono che una singola entità di tipo array venga trattata nel programma come l’insieme di tanti sotto-array, ciascuno dei quali può essere fisicamente allocato su una risorsa di memoria differente. Dal punto di vista del programma, indirizzare un array partizionato richiede che ad ogni accesso vengano eseguite delle istruzioni per ri-calcolare l’indirizzo fisico di destinazione. Questo è chiaramente un compito lungo, complesso e soggetto ad errori. Per questo motivo, le nostre tecniche di partizionamento sono state integrate nella l’interfaccia di programmazione di OpenMP, che è stata significativamente estesa. Specificamente, delle nuove direttive e clausole consentono al programmatore di annotare i dati di tipo array che si vuole partizionare e allocare in maniera distribuita sulla gerarchia di memoria. Sono stati inoltre sviluppati degli strumenti di supporto che consentono di raccogliere informazioni di profiling sul pattern di accesso agli array. Queste informazioni vengono sfruttate dal nostro compilatore per allocare le partizioni sulle varie risorse di memoria rispettando una relazione di affinità tra il task e i dati. Più precisamente, i passi di allocazione nel nostro compilatore assegnano una determinata partizione alla memoria scratchpad locale al processore che ospita il task che effettua il numero maggiore di accessi alla stessa.