High resolution sensors such as automotive radar and LiDAR have become prevalent in target tracking applications in recent times. Data from such sensors demands extended target tracking in which, the shape of the target is to be estimated along with the kinematics. Several applications benefit from extended target tracking, for example, autonomous vehicles and robotics. This thesis proposes a different approach to extended target tracking compared to existing literature. Instead of a single shape descriptor to describe the entire target shape, different parts of the extended target are assumed to be distinct targets constrained by the target rigid body shape. This formulation is able to handle issues such as self-occlusion and clutter which, are not addressed sufficiently in literature. Firstly, a framework for extended target tracking is developed based on the formulation proposed. Using 2D convex hull as a shape descriptor, an algorithm to track 2D convex polytope shaped targets is developed. Further, the point target Probabilistic Multiple Hypotheses Tracker (PMHT) is modified to derive an extended target PMHT (ET-PMHT) equations to track 3D convex polytope shapes, using a Delaunay triangulation to describe the shape. Finally, the approach is extended to handle target maneuvers, as well as, clutter and measurements from the interior of the target. In all three cases, the issue of self-occlusion is considered and the algorithms are still able to effectively capture the target shape. Since the true target center may not be observable, the shape descriptor abandons the use of target center in the state, and the shape is described by its boundary alone. The shape descriptors also support addition and deletion of faces, which is useful for handling newly visible parts of the target and clutter, respectively. The algorithms proposed have been compared with the existing literature for various scenarios, and it is seen that the proposed algorithms outperform, especially in the presence of self-occlusion. / Thesis / Candidate in Philosophy

Les réseaux de télécommunication doivent être fiables et robustes pour garantir la haute disponibilité des services. Les opérateurs cherchent actuellement à automatiser autant que possible les opérations complexes de gestion des réseaux, telles que le diagnostic de pannes.Dans cette thèse nous nous sommes intéressés au diagnostic automatique de pannes dans les réseaux d'accès optiques de l'opérateur Orange. L'outil de diagnostic utilisé jusqu'à présent, nommé DELC, est un système expert à base de règles de décision. Ce système est performant mais difficile à maintenir en raison, en particulier, du très grand volume d'informations à analyser. Il est également impossible de disposer d'une règle pour chaque configuration possible de panne, de sorte que certaines pannes ne sont actuellement pas diagnostiquées.Dans cette thèse nous avons proposé une nouvelle approche. Dans notre approche, le diagnostic des causes racines des anomalies et alarmes observées s'appuie sur une modélisation probabiliste, de type réseau bayésien, des relations de dépendance entre les différentes alarmes, compteurs, pannes intermédiaires et causes racines au niveau des différents équipements de réseau. Ce modèle probabiliste a été conçu de manière modulaire, de façon à pouvoir évoluer en cas de modification de l'architecture physique du réseau.Le diagnostic des causes racines des anomalies est effectué par inférence, dans le réseau bayésien, de l'état des noeuds non observés au vu des observations (compteurs, alarmes intermédiaires, etc...) récoltées sur le réseau de l'opérateur. La structure du réseau bayésien, ainsi que l'ordre de grandeur des paramètres probabilistes de ce modèle, ont été déterminés en intégrant dans le modèle les connaissances des experts spécialistes du diagnostic sur ce segment de réseau. L'analyse de milliers de cas de diagnostic de pannes a ensuite permis de calibrer finement les paramètres probabilistes du modèle grâce à un algorithme EM (Expectation Maximization).Les performances de l'outil développé, nommé PANDA, ont été évaluées sur deux mois de diagnostic de panne dans le réseau GPON-FTTH d'Orange en juillet-août 2015. Dans la plupart des cas, le nouveau système, PANDA, et le système en production, DELC, font un diagnostic identique. Cependant un certain nombre de cas sont non diagnostiqués par DELC mais ils sont correctement diagnostiqués par PANDA. Les cas pour lesquels les deux systèmes émettent des diagnostics différents ont été évalués manuellement, ce qui a permis de démontrer dans chacun de ces cas la pertinence des décisions prises par PANDA. / Telecommunication networks must be reliable and robust to ensure high availability of services. Operators are currently searching to automate as much as possible, complex network management operations such as fault diagnosis.In this thesis we are focused on self-diagnosis of failures in the optical access networks of the operator Orange. The diagnostic tool used up to now, called DELC, is an expert system based on decision rules. This system is efficient but difficult to maintain due in particular to the very large volume of information to analyze. It is also impossible to have a rule for each possible fault configuration, so that some faults are currently not diagnosed.We proposed in this thesis a new approach. In our approach, the diagnosis of the root causes of malfunctions and alarms is based on a Bayesian network probabilistic model of dependency relationships between the different alarms, counters, intermediate faults and root causes at the level of the various network component. This probabilistic model has been designed in a modular way, so as to be able to evolve in case of modification of the physical architecture of the network. Self-diagnosis of the root causes of malfunctions and alarms is made by inference in the Bayesian network model of the state of the nodes not observed in view of observations (counters, alarms, etc.) collected on the operator's network. The structure of the Bayesian network, as well as the order of magnitude of the probabilistic parameters of this model, were determined by integrating in the model the expert knowledge of the diagnostic experts on this segment of the network. The analysis of thousands of cases of fault diagnosis allowed to fine-tune the probabilistic parameters of the model thanks to an Expectation Maximization algorithm. The performance of the developed probabilistic tool, named PANDA, was evaluated over two months of fault diagnosis in Orange's GPON-FTTH network in July-August 2015. In most cases, the new system, PANDA, and the system in production, DELC, make an identical diagnosis. However, a number of cases are not diagnosed by DELC but are correctly diagnosed by PANDA. The cases for which self-diagnosis results of the two systems are different were evaluated manually, which made it possible to demonstrate in each of these cases the relevance of the decisions taken by PANDA.

In Natural Language Processing (NLP), speech and text are parsed and generated with language models and parser models, and translated with translation models. Each model contains a set of numerical parameters which are found by applying a suitable training algorithm to a set of training data. Many such training algorithms are instances of the Expectation-Maximization (EM) algorithm. In [BSV15], a generic EM algorithm for NLP is described. This work presents a particular speech model, the Hidden Markov model, and its standard training algorithm, the Baum-Welch algorithm. It is then shown that the Baum-Welch algorithm is an instance of the generic EM algorithm introduced by [BSV15], from which follows that all statements about the generic EM algorithm also apply to the Baum-Welch algorithm, especially its correctness and convergence properties.:1 Introduction 1.1 N-gram models 1.2 Hidden Markov model 2 Expectation-maximization algorithms 2.1 Preliminaries 2.2 Algorithmic skeleton 2.3 Corpus-based step mapping 2.4 Simple counting step mapping 2.5 Regular tree grammars 2.6 Inside-outside step mapping 2.7 Review 3 The Hidden Markov model 3.1 Forward and backward algorithms 3.2 The Baum-Welch algorithm 3.3 Deriving the Baum-Welch algorithm 3.3.1 Model parameter and countable events 3.3.2 Tree-shaped hidden information 3.3.3 Complete-data corpus 3.3.4 Inside weights 3.3.5 Outside weights 3.3.6 Complete-data corpus (cont.) 3.3.7 Step mapping 3.4 Review Appendix A Elided proofs from Chapter 3 A.1 Proof of Lemma 3.8 A.2 Proof of Lemma 3.9 B Formulary for Chapter 3 Bibliography

A phylogenetic tree is used to model gene evolution and species evolution using molecular sequence data. For artifactual and biological reasons, a gene tree may differ from a species tree, a phenomenon known as gene tree-species tree incongruence. Assuming the presence of one or more evolutionary events, e.g., gene duplication, gene loss, and lateral gene transfer (LGT), the incongruence may be explained using a reconciliation of a gene tree inside a species tree. Such information has biological utilities, e.g., inference of orthologous relationship between genes. In this thesis, we present probabilistic models and methods for orthology analysis and species tree inference, while accounting for evolutionary factors such as gene duplication, gene loss, and sequence evolution. Furthermore, we use a probabilistic LGT-aware model for inferring gene trees having temporal information for duplication and LGT events. In the first project, we present a Bayesian method, called DLRSOrthology, for estimating orthology probabilities using the DLRS model: a probabilistic model integrating gene evolution, a relaxed molecular clock for substitution rates, and sequence evolution. We devise a dynamic programming algorithm for efficiently summing orthology probabilities over all reconciliations of a gene tree inside a species tree. Furthermore, we present heuristics based on receiver operating characteristics (ROC) curve to estimate suitable thresholds for deciding orthology events. Our method, as demonstrated by synthetic and biological results, outperforms existing probabilistic approaches in accuracy and is robust to incomplete taxon sampling artifacts. In the second project, we present a probabilistic method, based on a mixture model, for species tree inference. The method employs a two-phase approach, where in the first phase, a structural expectation maximization algorithm, based on a mixture model, is used to reconstruct a maximum likelihood set of candidate species trees. In the second phase, in order to select the best species tree, each of the candidate species tree is evaluated using PrIME-DLRS: a method based on the DLRS model. The method is accurate, efficient, and scalable when compared to a recent probabilistic species tree inference method called PHYLDOG. We observe that, in most cases, the analysis constituted only by the first phase may also be used for selecting the target species tree, yielding a fast and accurate method for larger datasets. Finally, we devise a probabilistic method based on the DLTRS model: an extension of the DLRS model to include LGT events, for sampling reconciliations of a gene tree inside a species tree. The method enables us to estimate gene trees having temporal information for duplication and LGT events. To the best of our knowledge, this is the first probabilistic method that takes gene sequence data directly into account for sampling reconciliations that contains information about LGT events. Based on the synthetic data analysis, we believe that the method has the potential to identify LGT highways. / <p>QC 20150529</p>

Longitudinal data occur frequently in medical studies and covariates measured by error are typical features of such data. Generalized linear mixed models (GLMMs) are commonly used to analyse longitudinal data. It is typically assumed that the random effects covariance matrix is constant across the subject (and among subjects) in these models. In many situations, however, this correlation structure may differ among subjects and ignoring this heterogeneity can cause the biased estimates of model parameters. In this thesis, following Lee et al. (2012), we propose an approach to properly model the random effects covariance matrix based on covariates in the class of GLMMs where we also have covariates measured by error. The resulting parameters from this decomposition have a sensible interpretation and can easily be modelled without the concern of positive definiteness of the resulting estimator. The performance of the proposed approach is evaluated through simulation studies which show that the proposed method performs very well in terms biases and mean square errors as well as coverage rates. The proposed method is also analysed using a data from Manitoba Follow-up Study. / February 2017

The goal of this thesis research was to develop software to be used with RNA-Seq data for transcriptome quantification that was capable of handling multireads and quantifying isoforms on a more global level. Current software available for these purposes uses various forms of parameter alteration in order to work with multireads. Many still analyze isoforms per gene or per researcher determined clusters as well. By doing so, the effects of multireads are diminished or possibly wrongly represented. To address this issue, two programs, GWIE and ChromIE, were developed based on a simple iterative EM-like algorithm with no parameter manipulation. These programs are used to produce accurate isoform expression levels.

Continuous recognition of sign language has many practical applications and it can help to improve the quality of life of deaf persons by facilitating their interaction with hearing populace in public situations. This has led to some research in automated continuous American Sign Language recognition. But most work in continuous ASL recognition has only used top-down Hidden Markov Model (HMM) based approaches for recognition. There is no work on using facial information, which is considered to be fairly important. In this thesis, we explore bottom-up approach based on the use of Relational Distributions and Space of Probability Functions (SoPF) for intermediate level ASL recognition. We also use non-manual information, firstly, to decrease the number of deletion and insertion errors and secondly, to find whether the ASL sentence has 'Negation' in it, for which we use motion trajectories of the face. The experimental results show: The SoPF representation works well for ASL recognition. The accuracy based on the number of deletion errors, considering the 8 most probable signs in the sentence is 95%, while when considering 6 most probable signs, is 88%. Using facial or non-manual information increases accuracy when we consider top 6 signs, from 88% to 92%. Thus face does have information content in it. It is difficult to directly combine the manual information (information from hand motion) with non-manual (facial information) to improve the accuracy because of following two reasons: Manual images are not synchronized with the non-manual images. For example the same facial expressions is not present at the same manual position in two instances of the same sentences. One another problem in finding the facial expresion related with the sign, occurs when there is presence of a strong non-manual indicating 'Assertion' or 'Negation' in the sentence. In such cases the facial expressions are totally dominated by the face movements which is indicated by 'head shakes' or 'head nods'. The number of sentences, that have 'Negation' in them and are correctly recognized with the help of motion trajectories of the face are, 27 out of 30.

In this master's thesis, hidden Markov models (HMM) are evaluated as a tool for forecasting movements in a currency cross. With an ever increasing electronic market, making way for more automated trading, or so called algorithmic trading, there is constantly a need for new trading strategies trying to find alpha, the excess return, in the market. HMMs are based on the well-known theories of Markov chains, but where the states are assumed hidden, governing some observable output. HMMs have mainly been used for speech recognition and communication systems, but have lately also been utilized on financial time series with encouraging results. Both discrete and continuous versions of the model will be tested, as well as single- and multivariate input data. In addition to the basic framework, two extensions are implemented in the belief that they will further improve the prediction capabilities of the HMM. The first is a Gaussian mixture model (GMM), where one for each state assign a set of single Gaussians that are weighted together to replicate the density function of the stochastic process. This opens up for modeling non-normal distributions, which is often assumed for foreign exchange data. The second is an exponentially weighted expectation maximization (EWEM) algorithm, which takes time attenuation in consideration when re-estimating the parameters of the model. This allows for keeping old trends in mind while more recent patterns at the same time are given more attention. Empirical results shows that the HMM using continuous emission probabilities can, for some model settings, generate acceptable returns with Sharpe ratios well over one, whilst the discrete in general performs poorly. The GMM therefore seems to be an highly needed complement to the HMM for functionality. The EWEM however does not improve results as one might have expected. Our general impression is that the predictor using HMMs that we have developed and tested is too unstable to be taken in as a trading tool on foreign exchange data, with too many factors influencing the results. More research and development is called for.

This thesis deals with estimation of states and parameters in nonlinear and non-Gaussian dynamic systems. Sequential Monte Carlo methods are mainly used to this end. These methods rely on models of the underlying system, motivating some developments of the model concept. One of the main reasons for the interest in nonlinear estimation is that problems of this kind arise naturally in many important applications. Several applications of nonlinear estimation are studied. The models most commonly used for estimation are based on stochastic difference equations, referred to as state-space models. This thesis is mainly concerned with models of this kind. However, there will be a brief digression from this, in the treatment of the mathematically more intricate differential-algebraic equations. Here, the purpose is to write these equations in a form suitable for statistical signal processing. The nonlinear state estimation problem is addressed using sequential Monte Carlo methods, commonly referred to as particle methods. When there is a linear sub-structure inherent in the underlying model, this can be exploited by the powerful combination of the particle filter and the Kalman filter, presented by the marginalized particle filter. This algorithm is also known as the Rao-Blackwellized particle filter and it is thoroughly derived and explained in conjunction with a rather general class of mixed linear/nonlinear state-space models. Models of this type are often used in studying positioning and target tracking applications. This is illustrated using several examples from the automotive and the aircraft industry. Furthermore, the computational complexity of the marginalized particle filter is analyzed. The parameter estimation problem is addressed for a relatively general class of mixed linear/nonlinear state-space models. The expectation maximization algorithm is used to calculate parameter estimates from batch data. In devising this algorithm, the need to solve a nonlinear smoothing problem arises, which is handled using a particle smoother. The use of the marginalized particle filter for recursive parameterestimation is also investigated. The applications considered are the camera positioning problem arising from augmented reality and sensor fusion problems originating from automotive active safety systems. The use of vision measurements in the estimation problem is central to both applications. In augmented reality, the estimates of the camera’s position and orientation are imperative in the process of overlaying computer generated objects onto the live video stream. The objective in the sensor fusion problems arising in automotive safety systems is to provide information about the host vehicle and its surroundings, such as the position of other vehicles and the road geometry. Information of this kind is crucial for many systems, such as adaptive cruise control, collision avoidance and lane guidance.

<p>Landminor är ett stort problem både under och efter krigstid. De metoder som används för att detektera minor har inte ändrats mycket sedan 1940-talet. Forskning med mål att utvärdera olika elektro-optiska sensorer och metoder som skulle kunna användas för att skapa mer effektiv min-detektion genomförs på FOI. Försök som har gjorts med data från bland annat laser-radar och IR-sensorer har gett intressanta resultat.</p><p>I det här examensarbetet utvärderades olika fenomen och egenskaper i laser-radar- och IR-data. De testade egenskaperna var intensitet, IR, ytlikhet och höjd.</p><p>En metod som segmenterar intressanta objekt och bakgrundsdata utformades och implementerades. Metoden använde sig av expectation-maximization-skattning och ett minimum message length-kriterium. Ett scatter separability-kriterium användes för att bestämma kvalitén på de olika egenskaperna och på den resulterande segmenteringen.</p><p>Data insamlad under en mätkampanj av FOI användes för att testa metoden. Resultatet visade bland annat att ytlikhetsmåttet gav en bra segmentering för stora objekt med släta ytor, men var sämre för små objekt med skrovliga ytor. Vid jämförelse med en manuellt skapad mål-mask visade det sig att metoden klarade av att välja ut egenskaper som i många fall gav en godkänd segmentering.</p> / <p>Land mines are a huge problem in conflict time and after. Methods used to detect mines have not changed much since the 1940's. Research aiming to evaluate output from different electro-optical sensors and develop methods for more efficient mine detection is performed at FOI. Early experiments with laser radar sensors show promising results, as do analysis of data from infrared sensors.</p><p>In this thesis, an evaluation is made of features found in laser radar- and in infrared -sensor data. The tested features are intensity, infrared, a surfaceness feature extracted from the laser radar data and height above an estimated ground plane.</p><p>A method for segmenting interesting objects from background data using theexpectation-maximization algorithm and a minimum message length criterion is designed and implemented. A scatter separability criterion is utilized to determine the quality of the features and the resulting segmentation.</p><p>The method is tested on real data from a field trial performed by FOI. The results show that the surfaceness feature supports the segmentation of larger object with smooth surfaces but gives no contribution to small object with irregular surfaces. The method produces a decent result of selecting contributing features for different neighbourhoods of a scene. A comparison with a manually created target mask of the neighbourhood and the segmented components show that in most cases a high percentage separation of mine data and background data is possible.</p>