Multi Processor Systems On Chip with Configurable Hardware Acceleration

The evolution of the electronics embedded applications forces electronics systems designers to match their ever increasing requirements. This evolution pushes the computational power of digital signal processing systems, as well as the energy required to accomplish the computations, due to the increasing mobility of such applications. Current approaches used to match these requirements relies on the adoption of application specific signal processors. Such kind of devices exploits powerful accelerators, which are able to match both performance and energy requirements. On the other hand, the too high specificity of such accelerators often results in a lack of flexibility which affects non-recurrent engineering costs, time to market, and market volumes too. The state of the art mainly proposes two solutions to overcome these issues with the ambition of delivering reasonable performance and energy efficiency: reconfigurable computing and multi-processors computing. All of these solutions benefits from the post-fabrication programmability, that definitively results in an increased flexibility. Nevertheless, the gap between these approaches and dedicated hardware is still too high for many application domains, especially when targeting the mobile world. In this scenario, flexible and energy efficient acceleration can be achieved by merging these two computational paradigms, in order to address all the above introduced constraints. This thesis focuses on the exploration of the design and application spectrum of reconfigurable computing, exploited as application specific accelerators for multi-processors systems on chip. More specifically, it introduces a reconfigurable digital signal processor featuring a heterogeneous set of reconfigurable engines, and a homogeneous multi-core system, exploiting three different flavours of reconfigurable and mask-programmable technologies as implementation platform for applications specific accelerators. In this work, the various trade-offs concerning the utilization multi-core platforms and the different configuration technologies are explored, characterizing the design space of the proposed approach in terms of programmability, performance, energy efficiency and manufacturing costs.

Machine-learning methods for structure prediction of β-barrel membrane proteins

Different types of proteins exist with diverse functions that are essential for living organisms. An important class of proteins is represented by transmembrane proteins which are specifically designed to be inserted into biological membranes and devised to perform very important functions in the cell such as cell communication and active transport across the membrane. Transmembrane β-barrels (TMBBs) are a sub-class of membrane proteins largely under-represented in structure databases because of the extreme difficulty in experimental structure determination. For this reason, computational tools that are able to predict the structure of TMBBs are needed. In this thesis, two computational problems related to TMBBs were addressed: the detection of TMBBs in large datasets of proteins and the prediction of the topology of TMBB proteins. Firstly, a method for TMBB detection was presented based on a novel neural network framework for variable-length sequence classification. The proposed approach was validated on a non-redundant dataset of proteins. Furthermore, we carried-out genome-wide detection using the entire Escherichia coli proteome. In both experiments, the method significantly outperformed other existing state-of-the-art approaches, reaching very high PPV (92%) and MCC (0.82). Secondly, a method was also introduced for TMBB topology prediction. The proposed approach is based on grammatical modelling and probabilistic discriminative models for sequence data labeling. The method was evaluated using a newly generated dataset of 38 TMBB proteins obtained from high-resolution data in the PDB. Results have shown that the model is able to correctly predict topologies of 25 out of 38 protein chains in the dataset. When tested on previously released datasets, the performances of the proposed approach were measured as comparable or superior to the current state-of-the-art of TMBB topology prediction.

Pervasive Sensing in Future Networks

Pervasive Sensing is a recent research trend that aims at providing widespread computing and sensing capabilities to enable the creation of smart environments that can sense, process, and act by considering input coming from both people and devices. The capabilities necessary for Pervasive Sensing are nowadays available on a plethora of devices, from embedded devices to PCs and smartphones. The wide availability of new devices and the large amount of data they can access enable a wide range of novel services in different areas, spanning from simple data collection systems to socially-aware collaborative filtering. However, the strong heterogeneity and unreliability of devices and sensors poses significant challenges. So far, existing works on Pervasive Sensing have focused only on limited portions of the whole stack of available devices and data that they can use, to propose and develop mainly vertical solutions. The push from academia and industry for this kind of services shows that time is mature for a more general support framework for Pervasive Sensing solutions able to enhance frail architectures, promote a well balanced usage of resources on different devices, and enable the widest possible access to sensed data, while ensuring a minimal energy consumption on battery-operated devices. This thesis focuses on pervasive sensing systems to extract design guidelines as foundation of a comprehensive reference model for multi-tier Pervasive Sensing applications. The validity of the proposed model is tested in five different scenarios that present peculiar and different requirements, and different hardware and sensors. The ease of mapping from the proposed logical model to the real implementations and the positive performance result campaigns prove the quality of the proposed approach and offer a reliable reference model, together with a direction for the design and deployment of future Pervasive Sensing applications.

Evaluation of energy level to be absorbed by tractor ROPS: Actual Tests, Simulation and Computation

A highly dangerous situations for tractor driver is the lateral rollover in operating conditions. Several accidents, involving tractor rollover, have indeed been encountered, requiring the design of a robust Roll-Over Protective Structure (ROPS). The aim of the thesis was to evaluate tractor behaviour in the rollover phase so as to calculate the energy absorbed by the ROPS to ensure driver safety. A Mathematical Model representing the behaviour of a generic tractor during a lateral rollover, with the possibility of modifying the geometry, the inertia of the tractor and the environmental boundary conditions, is proposed. The purpose is to define a method allowing the prediction of the elasto-plastic behaviour of the subsequent impacts occurring in the rollover phase. A tyre impact model capable of analysing the influence of the wheels on the energy to be absorbed by the ROPS has been also developed. Different tractor design parameters affecting the rollover behaviour, such as mass and dimensions, have been considered. This permitted the evaluation of their influence on the amount of energy to be absorbed by the ROPS. The mathematical model was designed and calibrated with respect to the results of actual lateral upset tests carried out on a narrow-track tractor. The dynamic behaviour of the tractor and the energy absorbed by the ROPS, obtained from the actual tests, showed to match the results of the model developed. The proposed approach represents a valuable tool in understanding the dynamics (kinetic energy) and kinematics (position, velocity, angular velocity, etc.) of the tractor in the phases of lateral rollover and the factors mainly affecting the event. The prediction of the amount of energy to be absorbed in some cases of accident is possible with good accuracy. It can then help in designing protective structures or active security devices.

A new information fusion method for land-use classification using high resolution satellite imagery

Satellite image classification involves designing and developing efficient image classifiers. With satellite image data and image analysis methods multiplying rapidly, selecting the right mix of data sources and data analysis approaches has become critical to the generation of quality land-use maps. In this study, a new postprocessing information fusion algorithm for the extraction and representation of land-use information based on high-resolution satellite imagery is presented. This approach can produce land-use maps with sharp interregional boundaries and homogeneous regions. The proposed approach is conducted in five steps. First, a GIS layer - ATKIS data - was used to generate two coarse homogeneous regions, i.e. urban and rural areas. Second, a thematic (class) map was generated by use of a hybrid spectral classifier combining Gaussian Maximum Likelihood algorithm (GML) and ISODATA classifier. Third, a probabilistic relaxation algorithm was performed on the thematic map, resulting in a smoothed thematic map. Fourth, edge detection and edge thinning techniques were used to generate a contour map with pixel-width interclass boundaries. Fifth, the contour map was superimposed on the thematic map by use of a region-growing algorithm with the contour map and the smoothed thematic map as two constraints. For the operation of the proposed method, a software package is developed using programming language C. This software package comprises the GML algorithm, a probabilistic relaxation algorithm, TBL edge detector, an edge thresholding algorithm, a fast parallel thinning algorithm, and a region-growing information fusion algorithm. The county of Landau of the State Rheinland-Pfalz, Germany was selected as a test site. The high-resolution IRS-1C imagery was used as the principal input data.

Assessment of the resilience of coral reefs to natural and human disturbances by means of recruitment panels in Indo-Pacific

Indo-Pacific region encompasses about 75% of world's coral reefs, but hard coral cover in this region experienced a 32% region-wide decline since 1970s. This great change is primarily ascribable to natural and anthropogenic pressures, including climate change and human activities effects. Coral reef conservation requires management strategies oriented to maintain their diversity and the capacity to provide ecosystem goods and services. Coral reef resilience, i.e. the capacity to recover after disturbances, is critical to their long-term persistence. The aims of the present study were to design and to test field experiments intended to measure changes in recruitment processes, as a fundamental aspect of the coral reef resilience. Recruitment experiments, using artificial panels suspended in the water column, were carried out in two Indo-Pacific locations affected by different disturbances: a new mine in Bangka Island (Indonesia), and the increased sedimentation due to coastal dynamics in Vavvaru Island (Maldives). One (or more) putatively disturbed site(s) was selected to be tested against 3 randomly selected control sites. Panels’ arrangement simulates 2 proximities to living corals, i.e. the sources of propagules: few centimetres and 2 meters over. Panels were deployed simultaneously at each site and left submerged for about five months. Recruits were identified to the lowest possible taxonomic level and recruited assemblages were analysed in terms of percent cover. In general it was not possible to detect significant differences between the benthic assemblages recruited in disturbed and control sites. The high variability observed in recruits assemblages structure among control sites may be so large to mask the possible disturbance effects. Only few taxa showed possible effects of the disturb they undergo. The field tests have highlighted strengths and weaknesses of the proposed approach and, based on these results, some possible improvements were suggested.

Validity of an automatic measure protocol in distal femur for allograft selection from a three-dimensional virtual bone bank system

Osteoarticular allograft is one possible treatment in wide surgical resections with large defects. Performing best osteoarticular allograft selection is of great relevance for optimal exploitation of the bone databank, good surgery outcome and patient’s recovery. Current approaches are, however, very time consuming hindering these points in practice. We present a validation study of a software able to perform automatic bone measurements used to automatically assess the distal femur sizes across a databank. 170 distal femur surfaces were reconstructed from CT data and measured manually using a size measure protocol taking into account the transepicondyler distance (A), anterior-posterior distance in medial condyle (B) and anterior-posterior distance in lateral condyle (C). Intra- and inter-observer studies were conducted and regarded as ground truth measurements. Manual and automatic measures were compared. For the automatic measurements, the correlation coefficients between observer one and automatic method, were of 0.99 for A measure and 0.96 for B and C measures. The average time needed to perform the measurements was of 16 h for both manual measurements, and of 3 min for the automatic method. Results demonstrate the high reliability and, most importantly, high repeatability of the proposed approach, and considerable speed-up on the planning.

A Data-Analytic Strategy for Protein-Biomarker Discovery: Profiling of High-Dimensional Proteomic Data for Cancer Detection

With recent advances in mass spectrometry techniques, it is now possible to investigate proteins over a wide range of molecular weights in small biological specimens. This advance has generated data-analytic challenges in proteomics, similar to those created by microarray technologies in genetics, namely, discovery of "signature" protein profiles specific to each pathologic state (e.g., normal vs. cancer) or differential profiles between experimental conditions (e.g., treated by a drug of interest vs. untreated) from high-dimensional data. We propose a data analytic strategy for discovering protein biomarkers based on such high-dimensional mass-spectrometry data. A real biomarker-discovery project on prostate cancer is taken as a concrete example throughout the paper: the project aims to identify proteins in serum that distinguish cancer, benign hyperplasia, and normal states of prostate using the Surface Enhanced Laser Desorption/Ionization (SELDI) technology, a recently developed mass spectrometry technique. Our data analytic strategy takes properties of the SELDI mass-spectrometer into account: the SELDI output of a specimen contains about 48,000 (x, y) points where x is the protein mass divided by the number of charges introduced by ionization and y is the protein intensity of the corresponding mass per charge value, x, in that specimen. Given high coefficients of variation and other characteristics of protein intensity measures (y values), we reduce the measures of protein intensities to a set of binary variables that indicate peaks in the y-axis direction in the nearest neighborhoods of each mass per charge point in the x-axis direction. We then account for a shifting (measurement error) problem of the x-axis in SELDI output. After these pre-analysis processing of data, we combine the binary predictors to generate classification rules for cancer, benign hyperplasia, and normal states of prostate. Our approach is to apply the boosting algorithm to select binary predictors and construct a summary classifier. We empirically evaluate sensitivity and specificity of the resulting summary classifiers with a test dataset that is independent from the training dataset used to construct the summary classifiers. The proposed method performed nearly perfectly in distinguishing cancer and benign hyperplasia from normal. In the classification of cancer vs. benign hyperplasia, however, an appreciable proportion of the benign specimens were classified incorrectly as cancer. We discuss practical issues associated with our proposed approach to the analysis of SELDI output and its application in cancer biomarker discovery.

Use of multislice computed tomography in disaster victim identification--advantages and limitations

After a mass fatality incident (MFI), all victims have to be rapidly and accurately identified for juridical reasons as well as for the relatives' sake. Since MFIs are often international in scope, Interpol has proposed standard disaster victim identification (DVI) procedures, which have been widely adopted by authorities and forensic experts. This study investigates how postmortem multislice computed tomography (MSCT) can contribute to the DVI process as proposed by Interpol. The Interpol postmortem (PM) form has been analyzed, and a number of items in sections D and E thereof have been postulated to be suitable for documentation by CT data. CT scans have then been performed on forensic cases. Interpretation of the reconstructed images showed that indeed much of the postmortem information required for identification can be gathered from CT data. Further advantages of the proposed approach concern the observer independent documentation, the possibility to reconstruct a variety of images a long time after the event, the possibility to distribute the work by transmitting CT data digitally, and the reduction of time and specialists needed at the disaster site. We conclude that MSCT may be used as a valuable screening tool in DVI in the future.

A new system for computer-aided preoperative planning and intraoperative navigation during corrective jaw surgery

A new system for computer-aided corrective surgery of the jaws has been developed and introduced clinically. It combines three-dimensional (3-D) surgical planning with conventional dental occlusion planning. The developed software allows simulating the surgical correction on virtual 3-D models of the facial skeleton generated from computed tomography (CT) scans. Surgery planning and simulation include dynamic cephalometry, semi-automatic mirroring, interactive cutting of bone and segment repositioning. By coupling the software with a tracking system and with the help of a special registration procedure, we are able to acquire dental occlusion plans from plaster model mounts. Upon completion of the surgical plan, the setup is used to manufacture positioning splints for intraoperative guidance. The system provides further intraoperative assistance with the help of a display showing jaw positions and 3-D positioning guides updated in real time during the surgical procedure. The proposed approach offers the advantages of 3-D visualization and tracking technology without sacrificing long-proven cast-based techniques for dental occlusion evaluation. The system has been applied on one patient. Throughout this procedure, we have experienced improved assessment of pathology, increased precision, and augmented control.

Memory resource balancing for virtualized computing

Virtualization has become a common abstraction layer in modern data centers. By multiplexing hardware resources into multiple virtual machines (VMs) and thus enabling several operating systems to run on the same physical platform simultaneously, it can effectively reduce power consumption and building size or improve security by isolating VMs. In a virtualized system, memory resource management plays a critical role in achieving high resource utilization and performance. Insufficient memory allocation to a VM will degrade its performance dramatically. On the contrary, over-allocation causes waste of memory resources. Meanwhile, a VM’s memory demand may vary significantly. As a result, effective memory resource management calls for a dynamic memory balancer, which, ideally, can adjust memory allocation in a timely manner for each VM based on their current memory demand and thus achieve the best memory utilization and the optimal overall performance. In order to estimate the memory demand of each VM and to arbitrate possible memory resource contention, a widely proposed approach is to construct an LRU-based miss ratio curve (MRC), which provides not only the current working set size (WSS) but also the correlation between performance and the target memory allocation size. Unfortunately, the cost of constructing an MRC is nontrivial. In this dissertation, we first present a low overhead LRU-based memory demand tracking scheme, which includes three orthogonal optimizations: AVL-based LRU organization, dynamic hot set sizing and intermittent memory tracking. Our evaluation results show that, for the whole SPEC CPU 2006 benchmark suite, after applying the three optimizing techniques, the mean overhead of MRC construction is lowered from 173% to only 2%. Based on current WSS, we then predict its trend in the near future and take different strategies for different prediction results. When there is a sufficient amount of physical memory on the host, it locally balances its memory resource for the VMs. Once the local memory resource is insufficient and the memory pressure is predicted to sustain for a sufficiently long time, a relatively expensive solution, VM live migration, is used to move one or more VMs from the hot host to other host(s). Finally, for transient memory pressure, a remote cache is used to alleviate the temporary performance penalty. Our experimental results show that this design achieves 49% center-wide speedup.

Synthesis of computations and experiments for obtaining pulsatile gas flow rates from dynamic pressure difference measurements across an orifice-plate meter

The use of conventional orifice-plate meter is typically restricted to measurements of steady flows. This study proposes a new and effective computational-experimental approach for measuring the time-varying (but steady-in-the-mean) nature of turbulent pulsatile gas flows. Low Mach number (effectively constant density) steady-in-the-mean gas flows with large amplitude fluctuations (whose highest significant frequency is characterized by the value fF) are termed pulsatile if the fluctuations have a direct correlation with the time-varying signature of the imposed dynamic pressure difference and, furthermore, they have fluctuation amplitudes that are significantly larger than those associated with turbulence or random acoustic wave signatures. The experimental aspect of the proposed calibration approach is based on use of Coriolis-meters (whose oscillating arm frequency fcoriolis >> fF) which are capable of effectively measuring the mean flow rate of the pulsatile flows. Together with the experimental measurements of the mean mass flow rate of these pulsatile flows, the computational approach presented here is shown to be effective in converting the dynamic pressure difference signal into the desired dynamic flow rate signal. The proposed approach is reliable because the time-varying flow rate predictions obtained for two different orifice-plate meters exhibit the approximately same qualitative, dominant features of the pulsatile flow.

Time-domain modeling of high-frequency electromagnetic wave propagation, overhead wires, and earth

Prediction of radiated fields from transmission lines has not previously been studied from a panoptical power system perspective. The application of BPL technologies to overhead transmission lines would benefit greatly from an ability to simulate real power system environments, not limited to the transmission lines themselves. Presently circuitbased transmission line models used by EMTP-type programs utilize Carson’s formula for a waveguide parallel to an interface. This formula is not valid for calculations at high frequencies, considering effects of earth return currents. This thesis explains the challenges of developing such improved models, explores an approach to combining circuit-based and electromagnetics modeling to predict radiated fields from transmission lines, exposes inadequacies of simulation tools, and suggests methods of extending the validity of transmission line models into very high frequency ranges. Electromagnetics programs are commonly used to study radiated fields from transmission lines. However, an approach is proposed here which is also able to incorporate the components of a power system through the combined use of EMTP-type models. Carson’s formulas address the series impedance of electrical conductors above and parallel to the earth. These equations have been analyzed to show their inherent assumptions and what the implications are. Additionally, the lack of validity into higher frequencies has been demonstrated, showing the need to replace Carson’s formulas for these types of studies. This body of work leads to several conclusions about the relatively new study of BPL. Foremost, there is a gap in modeling capabilities which has been bridged through integration of circuit-based and electromagnetics modeling, allowing more realistic prediction of BPL performance and radiated fields. The proposed approach is limited in its scope of validity due to the formulas used by EMTP-type software. To extend the range of validity, a new set of equations must be identified and implemented in the approach. Several potential methods of implementation have been explored. Though an appropriate set of equations has not yet been identified, further research in this area will benefit from a clear depiction of the next important steps and how they can be accomplished. Prediction of radiated fields from transmission lines has not previously been studied from a panoptical power system perspective. The application of BPL technologies to overhead transmission lines would benefit greatly from an ability to simulate real power system environments, not limited to the transmission lines themselves. Presently circuitbased transmission line models used by EMTP-type programs utilize Carson’s formula for a waveguide parallel to an interface. This formula is not valid for calculations at high frequencies, considering effects of earth return currents. This thesis explains the challenges of developing such improved models, explores an approach to combining circuit-based and electromagnetics modeling to predict radiated fields from transmission lines, exposes inadequacies of simulation tools, and suggests methods of extending the validity of transmission line models into very high frequency ranges. Electromagnetics programs are commonly used to study radiated fields from transmission lines. However, an approach is proposed here which is also able to incorporate the components of a power system through the combined use of EMTP-type models. Carson’s formulas address the series impedance of electrical conductors above and parallel to the earth. These equations have been analyzed to show their inherent assumptions and what the implications are. Additionally, the lack of validity into higher frequencies has been demonstrated, showing the need to replace Carson’s formulas for these types of studies. This body of work leads to several conclusions about the relatively new study of BPL. Foremost, there is a gap in modeling capabilities which has been bridged through integration of circuit-based and electromagnetics modeling, allowing more realistic prediction of BPL performance and radiated fields. The proposed approach is limited in its scope of validity due to the formulas used by EMTP-type software. To extend the range of validity, a new set of equations must be identified and implemented in the approach. Several potential methods of implementation have been explored. Though an appropriate set of equations has not yet been identified, further research in this area will benefit from a clear depiction of the next important steps and how they can be accomplished.

Computational simulations of latent heat thermal energy storage systems - with innovative and first-principles based simulation for the underlying unsteady melting (and solidification) processes

This thesis develops an effective modeling and simulation procedure for a specific thermal energy storage system commonly used and recommended for various applications (such as an auxiliary energy storage system for solar heating based Rankine cycle power plant). This thermal energy storage system transfers heat from a hot fluid (termed as heat transfer fluid - HTF) flowing in a tube to the surrounding phase change material (PCM). Through unsteady melting or freezing process, the PCM absorbs or releases thermal energy in the form of latent heat. Both scientific and engineering information is obtained by the proposed first-principle based modeling and simulation procedure. On the scientific side, the approach accurately tracks the moving melt-front (modeled as a sharp liquid-solid interface) and provides all necessary information about the time-varying heat-flow rates, temperature profiles, stored thermal energy, etc. On the engineering side, the proposed approach is unique in its ability to accurately solve – both individually and collectively – all the conjugate unsteady heat transfer problems for each of the components of the thermal storage system. This yields critical system level information on the various time-varying effectiveness and efficiency parameters for the thermal storage system.

Genetic association studies under the population stratification, family pedigree and application to genome-wide association studies

This dissertation has three separate parts: the first part deals with the general pedigree association testing incorporating continuous covariates; the second part deals with the association tests under population stratification using the conditional likelihood tests; the third part deals with the genome-wide association studies based on the real rheumatoid arthritis (RA) disease data sets from Genetic Analysis Workshop 16 (GAW16) problem 1. Many statistical tests are developed to test the linkage and association using either case-control status or phenotype covariates for family data structure, separately. Those univariate analyses might not use all the information coming from the family members in practical studies. On the other hand, the human complex disease do not have a clear inheritance pattern, there might exist the gene interactions or act independently. In part I, the new proposed approach MPDT is focused on how to use both the case control information as well as the phenotype covariates. This approach can be applied to detect multiple marker effects. Based on the two existing popular statistics in family studies for case-control and quantitative traits respectively, the new approach could be used in the simple family structure data set as well as general pedigree structure. The combined statistics are calculated using the two statistics; A permutation procedure is applied for assessing the p-value with adjustment from the Bonferroni for the multiple markers. We use simulation studies to evaluate the type I error rates and the powers of the proposed approach. Our results show that the combined test using both case-control information and phenotype covariates not only has the correct type I error rates but also is more powerful than the other existing methods. For multiple marker interactions, our proposed method is also very powerful. Selective genotyping is an economical strategy in detecting and mapping quantitative trait loci in the genetic dissection of complex disease. When the samples arise from different ethnic groups or an admixture population, all the existing selective genotyping methods may result in spurious association due to different ancestry distributions. The problem can be more serious when the sample size is large, a general requirement to obtain sufficient power to detect modest genetic effects for most complex traits. In part II, I describe a useful strategy in selective genotyping while population stratification is present. Our procedure used a principal component based approach to eliminate any effect of population stratification. The paper evaluates the performance of our procedure using both simulated data from an early study data sets and also the HapMap data sets in a variety of population admixture models generated from empirical data. There are one binary trait and two continuous traits in the rheumatoid arthritis dataset of Problem 1 in the Genetic Analysis Workshop 16 (GAW16): RA status, AntiCCP and IgM. To allow multiple traits, we suggest a set of SNP-level F statistics by the concept of multiple-correlation to measure the genetic association between multiple trait values and SNP-specific genotypic scores and obtain their null distributions. Hereby, we perform 6 genome-wide association analyses using the novel one- and two-stage approaches which are based on single, double and triple traits. Incorporating all these 6 analyses, we successfully validate the SNPs which have been identified to be responsible for rheumatoid arthritis in the literature and detect more disease susceptibility SNPs for follow-up studies in the future. Except for chromosome 13 and 18, each of the others is found to harbour susceptible genetic regions for rheumatoid arthritis or related diseases, i.e., lupus erythematosus. This topic is discussed in part III.