Microarray technology allows the simultaneous measurement of thousands of gene expressions simultaneously. As a result of this, many statistical methods emerged for identifying differentially expressed genes. Kerr et al. (2001) proposed analysis of variance (ANOVA) procedure for the analysis of gene expression data. Their estimators are based on the assumption of normality, however the parameter estimates and residuals from this analysis are notably heavier-tailed than normal as they commented. Since non-normality complicates the data analysis and results in inefficient estimators, it is very important to develop statistical procedures which are efficient and robust. For this reason, in this work, we use Modified Maximum Likelihood (MML) and Adaptive Maximum Likelihood estimation method (Tiku and Suresh, 1992) and show that MML and AMML estimators are more efficient and robust. In our study we compared MML and AMML method with widely used statistical analysis methods via simulations and real microarray data sets.

Data stream processing in the industrial as well as in the academic field has gained more and more importance during the last years. Consider the monitoring of industrial processes as an example. There, sensors are mounted to gather lots of data within a short time range. Storing and post-processing these data may occasionally be useless or even impossible. On the one hand, only a small part of the monitored data is relevant. To efficiently use the storage capacity, only a preselection of the data should be considered. On the other hand, it may occur that the volume of incoming data is generally too high to be stored in time or–in other words–the technical efforts for storing the data in time would be out of scale. Processing data streams in the context of this thesis means to apply database operations to the stream in an on-the-fly manner (without explicitly storing the data). The challenges for this task lie in the limited amount of resources while data streams are potentially infinite. Furthermore, data stream processing must be fast and the results have to be disseminated as soon as possible. This thesis focuses on the latter issue. The goal is to provide a so-called Quality-of-Service (QoS) for the data stream processing task. Therefore, adequate QoS metrics like maximum output delay or minimum result data rate are defined. Thereafter, a cost model for obtaining the required processing resources from the specified QoS is presented. On that basis, the stream processing operations are scheduled. Depending on the required QoS and on the available resources, the weight can be shifted among the individual resources and QoS metrics, respectively. Calculating and scheduling resources requires a lot of expert knowledge regarding the characteristics of the stream operations and regarding the incoming data streams. Often, this knowledge is based on experience and thus, a revision of the resource calculation and reservation becomes necessary from time to time. This leads to occasional interruptions of the continuous data stream processing, of the delivery of the result, and thus, of the negotiated Quality-of-Service. The proposed robustness concept supports the user and facilitates a decrease in the number of interruptions by providing more resources.

<p>In this thesis, the problem of controlling a surgical masterand slave system with force reflection is studied. The problemof stiff contacts between the slave and the environment isgiven specific attention. The work has been carried out at KTHbased on an initial cooperation with Karolinska Sjukhuset. Theaim of the over all project is to study the possibilities forintroduction of a force reflective teleoperator in neurologicalskullbase operations for the particular task of bone millingand thereby, hopefully, increase patient safety, decreasesurgeon workload and cost forthe society.</p><p>The main contributions of this thesis are:</p><p>Derivation of a dynamical model of the master andoperator’s finger system and, experimental identificationof ranges on model parameter values. Based on this model, theinteraction channel controllers optimized for transparency arederived and modified to avoid the influence of the uncertainmodel parameters. This results in a three channel structure. Todecrease the influence of the uncertain parameters locally atthe master, a control loop is designed such that the frequencyresponse of the reflected force is relatively unaffected by theuncertainties, a result also confirmed in a transparencyanalysis based on the H-matrix. The developed teleoperatorcontrol structure is tested in experiments where the operatorcould alter the contact force without facing any problems aslong as the slave is in contact with the environment.</p><p>As a result of the severe difficulties for the teleoperatorto move from free space motion to in-contact manipulationwithout oscillative behaviour, a new detection algorithm basedon passivity theory is developed. The algorithm is able todetect the non-passive behaviour of the actual teleoperatorinduced by the discrete change in system dynamics occurring atthe contact instant. A stabilization controller to be activatedby the detection algorithm is designed and implemented on themaster side of the teleoperator. The detection algorithm andthe stabilization controller are shown highly effective in realexperiments.</p><p>All major research results presented in the thesis have beenverified experimentally.</p><p><b>Keywords</b>Teleoperator, Force Feedback, Passivity, StiffContacts, Control, Robustness, Transparency, Bone Milling,Uncertainty</p>

<p>This thesis concerns optimization of structures considering various uncertainties. The overall objective is to find methods to create solutions that are optimal both in the sense of handling the typical load case and minimising the variability of the response, i.e. robust optimal designs.</p><p>Traditionally optimized structures may show a tendency of being sensitive to small perturbations in the design or loading conditions, which of course are inevitable. To create robust designs, it is necessary to account for all conceivable variations (or at least the influencing ones) in the design process.</p><p>The thesis is divided in two parts. The first part serves as a theoretical background to the second part, the two appended articles. This first part includes the concept of robust design, basic statistics, optimization theory and meta modelling.</p><p>The first appended paper is an application of existing methods on a large industrial example problem. A sensitivity analysis is performed on a Scania truck cab subjected to impact loading in order to identify the most influencing variables on the crash responses.</p><p>The second paper presents a new method that may be used in robust optimizations, that is, optimizations that account for variations and uncertainties. The method is demonstrated on both an analytical example and a Finite Element example of an aluminium extrusion subjected to axial crushing.</p> / ROBDES

Si la communauté internationale de géométrie algorithmique a souvent<br />la tentation de s'engouffrer dans des recherches essentiellement<br />théoriques, et en particulier combinatoires, la grande originalité des<br />travaux à l'INRIA résidait déjà à l'époque de mes débuts dans le<br />souci de leur validation expérimentale et de leur applicabilité. <br /><br />Le domaine a suivi globalement une évolution dans cette direction,<br />en particulier grâce à l'``Impact Task Force Report''. Notre intérêt pour le transfert technologique et<br />industriel, ainsi que pour l'établissement d'une plateforme pour la<br />recherche, a pris pendant ce temps une tournure encore plus concrète<br />avec notre implication très forte dans le projet CGAL<br />dont notre équipe est l'un des moteurs.<br /><br />Ce document prend le parti de présenter les travaux sous l'angle de<br />cette préoccupation pratique.<br />Il comporte deux chapitres principaux : le premier rassemble<br />des travaux sur les triangulations, le second présente des travaux sur<br />les objets courbes. Ces deux chapitres se concluent par un ensemble de<br />directions ouvertes. Le troisième chapitre survole rapidement d'autres<br />résultats.

It is a challenging problem to make a general hand gesture recognition system work in a practical operation environment. In this study, it is mainly focused on recognizing English letters and digits performed near the steering wheel of a car and captured by a video camera. Like most human computer interaction (HCI) scenarios, the in-car gesture recognition suffers from various robustness issues, including multiple human factors and highly varying lighting conditions. It therefore brings up quite a few research issues to be addressed. First, multiple gesturing alternatives may share the same meaning, which is not typical in most previous systems. Next, gestures may not be the same as expected because users cannot see what exactly has been written, which increases the gesture diversity significantly.In addition, varying illumination conditions will make hand detection trivial and thus result in noisy hand gestures. And most severely, users will tend to perform letters at a fast pace, which may result in lack of frames for well-describing gestures. Since users are allowed to perform gestures in free-style, multiple alternatives and variations should be considered while modeling gestures. The main contribution of this work is to analyze and address these challenging issues step-by-step such that eventually the robustness of the whole system can be effectively improved. By choosing color-space representation and performing the compensation techniques for varying recording conditions, the hand detection performance for multiple illumination conditions is first enhanced. Furthermore, the issues of low frame rate and different gesturing tempo will be separately resolved via the cubic B-spline interpolation and i-vector method for feature extraction. Finally, remaining issues will be handled by other modeling techniques such as sub-letter stroke modeling. According to experimental results based on the above strategies, the proposed framework clearly improved the system robustness and thus encouraged the future research direction on exploring more discriminative features and modeling techniques.

Software systems are ubiquitous in today's world and yet, remain vulnerable to the fallibility of human programmers as well as the unpredictability of their operating environments. The overarching goal of this dissertation is to develop algorithms to enable automated and efficient design and analysis of reliable programs. In the first and second parts of this dissertation, we focus on the development of programs that are free from programming errors. The intent is not to eliminate the human programmer, but instead to complement his or her expertise, with sound and efficient computational techniques, when possible. To this end, we make contributions in two specific domains. Program debugging --- the process of fault localization and error elimination from a program found to be incorrect --- typically relies on expert human intuition and experience, and is often a lengthy, expensive part of the program development cycle. In the first part of the dissertation, we target automated debugging of sequential programs. A broad and informal statement of the (automated) program debugging problem is to suitably modify an erroneous program, say P, to obtain a correct program, say P'. This problem is undecidable in general; it is hard to formalize; moreover, it is particularly challenging to assimilate and mechanize the customized, expert programmer intuition involved in the choices made in manual program debugging. Our first contribution in this domain is a methodical formalization of the program debugging problem, that enables automation, while incorporating expert programmer intuition and intent. Our second contribution is a solution framework that can debug infinite-state, imperative, sequential programs written in higher-level programming languages such as C. Boolean programs, which are smaller, finite-state abstractions of infinite-state or large, finite-state programs, have been found to be tractable for program verification. In this dissertation, we utilize Boolean programs for program debugging. Our solution framework involves two main steps: (a) automated debugging of a Boolean program, corresponding to an erroneous program P, and (b) translation of the corrected Boolean program into a correct program P'. Shared-memory concurrent programs are notoriously difficult to write, verify and debug; this makes them excellent targets for automated program completion, in particular, for synthesis of synchronization code. Extant work in this domain has focused on either propositional temporal logic specifications with simplistic models of concurrent programs, or more refined program models with the specifications limited to just safety properties. Moreover, there has been limited effort in developing adaptable and fully-automatic synthesis frameworks that are capable of generating synchronization at different levels of abstraction and granularity. In the second part of this dissertation, we present a framework for synthesis of synchronization for shared-memory concurrent programs with respect to temporal logic specifications. In particular, given a concurrent program composed of synchronization-free processes, and a temporal logic specification describing their expected concurrent behaviour, we generate synchronized processes such that the resulting concurrent program satisfies the specification. We provide the ability to synthesize readily-implementable synchronization code based on lower-level primitives such as locks and condition variables. We enable synchronization synthesis of finite-state concurrent programs composed of processes that may have local and shared variables, may be straight-line or branching programs, may be ongoing or terminating, and may have program-initialized or user-initialized variables. We also facilitate expression of safety and liveness properties over both control and data variables by proposing an extension of propositional computation tree logic. Most program analyses, verification, debugging and synthesis methodologies target traditional correctness properties such as safety and liveness. These techniques typically do not provide a quantitative measure of the sensitivity of a computational system's behaviour to unpredictability in the operating environment. We propose that the core property of interest in reasoning in the presence of such uncertainty is robustness --- small perturbations to the operating environment do not change the system's observable behavior substantially. In well-established areas such as control theory, robustness has always been a fundamental concern; however, the techniques and results therein are not directly applicable to computational systems with large amounts of discretized, discontinuous behavior. Hence, robustness analysis of software programs used in heterogeneous settings necessitates development of new theoretical frameworks and algorithms. In the third part of this dissertation, we target robustness analysis of two important classes of discrete systems --- string transducers and networked systems of Mealy machines. For each system, we formally define robustness of the system with respect to a specific source of uncertainty. In particular, we analyze the behaviour of transducers in the presence of input perturbations, and the behaviour of networked systems in the presence of channel perturbations. Our overall approach is automata-theoretic, and necessitates the use of specialized distance-tracking automata for tracking various distance metrics between two strings. We present constructions for such automata and use them to develop decision procedures based on reducing the problem of robustness verification of our systems to the problem of checking the emptiness of certain automata. Thus, the system under consideration is robust if and only if the languages of particular automata are empty. / text

Σκοπός της συγκεκριμένης εργασίας είναι η παρουσίαση των σημαντικότερων τεχνικών υδατογράφησης (watermarking techniques) ψηφιακών εικόνων με ανθεκτικότητα σε γεωμετρικούς μετασχηματισμούς και η περιγραφή και υλοποίηση ενός τέτοιου συστήματος. Οι γεωμετρικοί μετασχηματισμοί, στους οποίους δόθηκε περισσότερη έμφαση κατά την υλοποίηση του συστήματος υδατογράφησης ψηφιακών εικόνων που θα παρουσιαστεί παρακάτω, έχουν να κάνουν με την κλιμάκωση (scaling), την περιστροφή (rotation) και τον μετασχηματισμό συντεταγμένων (translation) μιας εικόνας καθώς επίσης και με συνδυασμούς αυτών. Παρόλα αυτά όμως το σύστημα που περιγράφεται αντιμετωπίζει σε ικανοποιητικό βαθμό και άλλες γεωμετρικές επιθέσεις (όπως shearing,aspect ratio change, projective, template remove κ.α.).Η ανθεκτικότητα ενός συστήματος υδατογράφησης σε αυτούς τους μετασχηματισμούς έχει να κάνει με την διατήρηση του υδατογραφήματος και την ικανοποιητική ανίχνευση του σε εικόνες που έχουν υποστεί τέτοιου είδους μετασχηματισμούς. Στην εργασία, αρχικά γίνεται μια εκτενής αναφορά στην τεχνική της υδατογράφησης και τις εφαρμογές της και έπειτα τονίζονται τα σημαντικότερα χαρακτηριστικά των συστημάτων υδατογράφησης ψηφιακού υλικού. Στη συνέχεια, ορίζονται οι γεωμετρικές επιθέσεις σε εικόνες και παρουσιάζονται οι σημαντικότερες μέθοδοι υδατογράφησης ψηφιακών εικόνων, που υπάρχουν στην βιβλιογραφία, με ανθεκτικότητα στις επιθέσεις αυτές. Στην εργασία αυτή υλοποιήθηκε μία από αυτές τις μεθόδους και έγινε προσπάθεια για επέκταση και βελτίωση της σε ορισμένα σημεία που υστερούσε. Η αρχική και βελτιωμένη μέθοδος περιγράφονται αναλυτικά. Στο τέλος ένας αριθμός πειραματικών αποτελεσμάτων αποδεικνύουν την ανθεκτικότητα του νέου βελτιωμένου συστήματος που υλοποιήθηκε σε ένα μεγάλο πλήθος γεωμετρικών επιθέσεων. / The aim of this work is the presentation of the most important watermarking techniques for digital images with robustness to geometric transformations and the description and implementation of such a system. The geometric transformations, to which was given more emphasis at the implementation of the digital images watermarking system, that will be presented below, are the scaling, the rotation and the transformation of coordinates (translation) of an image as well as the combinations of these. Nevertheless, the system that is described faces in satisfactory degree and the other geometric attacks (as the shearing, the aspect ratio change, the projective, the template remove etc.). The robustness of a watermarking system to these transformations it is simply translated as the maintenance of the watermark and its satisfactory detection in images that have existed such type of transformations. In this work, it initially becomes a extensive report to the watermark technique and to its applications and then the more important characteristics of digital material watermarking systems are highlighted. Afterwards, the geometric attacks of digit images are determined and the most important watermarking methods of digital images with robustness to these attacks, that exist in the bibliography, are presented. In this work, one of these methods was implemented and it became some effort for its extension and its improvement in certain points that disadvantages. The initial and improved methods are analytically described. At the end, a number of experimental results are presented, that they prove the robustness of the new improved system to a large number of geometric attacks.

Timed automata are a formalism to model, verify, and synthesize real-time systems. They have the advantage of having an abstract mathematical semantics, which allow formalizing and solving several verification and synthesis problems. However, timed automata are intended to design models, rather than completely describe real systems. Therefore, once the design phase is over, it remains to check whether the behavior of an actual implementation corresponds to that of the timed automaton model. An important step before implementing a system design is ensuring its robustness. This thesis considers a notion of robustness that asks whether the behavior of a given timed automaton is preserved, or can be made so, when it is subject to small perturbations. Several approaches are considered: Robustness analysis seeks to decide whether a given timed automaton tolerates perturbations, and in that case to compute the (maximum) amount of tolerated perturbations. In robust synthesis, a given system needs to be controlled by a law (or strategy) which tolerates perturbations upto some computable amount. In robust implementation, one seeks to automatically transform a given timed automaton model so that it tolerates perturbations by construction. Several perturbation models are considered, ranging from introducing error in time measures (guard enlargement), forbidding behaviors that are too close to boundaries (guard shrinking), and restricting the time domain to a discrete sampling. We also formalize robust synthesis problems as games, where the control law plays against the environment which can systematically perturb the chosen moves, by some bounded amount. These problems are studied on timed automata and their variants, namely, timed games, and weighted timed automata and games. Algorithms for the parameterized robustness analysis against guard enlargements, and guard shrinkings are presented. The robust synthesis problem is studied for two variants of the game semantics, for timed automata, games, and their weighted extensions. A software tool for robustness analysis against guard shrinkings is presented, and experimental results are discussed. The robust implementation problem is also studied in two different settings. In all algorithms, an upper bound on perturbations that the given timed automaton tolerates can be computed.

Simultaneous stabilization is an important problem in the design of robust controllers. It is the problem of designing a single feedback controller which will simultaneously stabilize every member of a finite collection of liner time-invariant systems. This provides simplicity and reliability which is desirable in aerospace applications. It can be used as a back-up control system in sophisticated airplanes, or an inexpensive primary one for small aircraft. In this work the robustness of the simultaneous stabilization problem, known as the Robust Simultaneous Stabilization (RSS) problem, is addressed. First, an optimization methodology for finding a solution to the Simultaneous Stabilization (SS) problem is proposed. Next, in order to provide simultaneous stability while maximizing the stability robustness bounds, a multiple-robustness optimization design methodology for the RSS problem is presented. The two proposed design methodologies are then compared in terms of robustness of the designed controller.