Technical systems of today are often complex and integrated. If a fault occurs, the consequences can be disastrous both for the system itself and its surroundings. To maintain the operation and the security it is necessary to have a surveillance system which can detect a fault in an early stage. In this thesis two structural methods for fault isolation analysis are discussed. The result from the studied algorithms shows what fault isolation properties a diagnostic model is expected to have. If the isolability is not good enough, it also gives information on where further modelling needs to be done. To base a comparison of the two structural analysis algorithms on, four criteria are defined concerning for example realizability of residuals and time complexity. One interesting part of the methods is how dynamic models are handled. It is shown how differential constraints can end up in differential cycles which implies calculatory problems and what effects structural differentiation has on a system. The algorithms have been tested on an application from the research training network DAMADICS. The result shows how different types of input models in this case give the same result.

Software testing is a vital part of the software development process, usually implemented at the coding stage, and costly. Due to the increased use of the Unified Modeling Language (UML) and Model Driven Architecture (MDA) approach in systems analysis and design, Model-Based Testing has been discussed as a prominent solution for software testing to reduce the cost of software testing. Prior researches proposed an integrated method which utilizes the artifacts from the Platform Independent Model (PIM) to construct the test paths and generate the test cases. This study develops a methodology which extracting the information from PIM (e.g., Sequence Diagram and Class diagram) to generate the test cases directly. The research methodology is articulated using the design science research methodology. A usability evaluation is performed to demonstrate its applicability. With this methodology, the test cases can be easily generated; thereby reducing the cost and enhancing the efficiency of Model-Based Testing.

Software testing is a vital part of the software development process and is costly. Due to the increased use of the unified modeling language and model driven architecture approach in systems analysis and design, model-based testing has been discussed as a prominent solution for software testing to reduce the cost of software testing. Prior researches proposed an integrated method which utilizes the artifacts from the Platform Independent Model (PIM) to construct the test path and constrained class tuples. These two can then be integrated to construct an Integrated Testing Model (ITM) for determining the test data and test cases for Model-Based Testing. This study develops a methodology which extracting the information from ITM to construct the test cases. The research methodology is articulated using the design science research methodology. A prototype embedded methodology has been developed to automatically generate the test cases. A usability evaluation is performed on the prototype to demonstrate its usability. With this methodology, the test cases can be generated automatically; thereby reducing the cost and enhancing the efficiency of Model-Based Testing.

<p>If a fault occurs in a technical system, for example in an airplane, it is important to beable to detect that there is a fault and to find what in the system that is faulty. Theprocedure of determining, given certain observations, if faults are present and if so thelocation of faults is called a diagnosis. For achieving diagnosis we can use computersoftware that takes observations of a system as input and that generates a diagnosis asoutput. This is called a diagnostic system. To build a diagnostic system we needanother piece of computer software which is called a diagnostic tool. This thesis willpresent a market survey for diagnostic tools as well as an analysis of three of the toolsfound in the survey. The analysis can be seen as constituted by two different aspects,one focusing on the diagnostic methods with which each tool creates diagnosticsystems, the other focusing on practical details that determine the usability of eachtool. The analysis found that the largest differences were between the methods used increating the diagnostic systems.</p>

The paper focuses on the design and implementation of the base-band basic receiving functions, for a binary CP-FSK demodulator pilot study, as independent modules of a complete Reconfigurable Data-Link (RDL). A model-based approach and Software Defined Radio (SDR) paradigm are used for the design. The implementation will be executed on Field-Programmable Gate Array (FPGA) based hardware.

Designing high-integrity socio-technical systems requires a thorough understanding of all safety risks of such systems. For many years, safety risk assessment has been conducted separately for hardware, software, human, organizational and other entities in socio-technical systems. Safety risk assessment that does not consider all factors at the same time cannot adequately capture the wide variety of safety risk scenarios that need to be considered. This thesis proposes a model-based analysis approach that allows interpretation of humans and organizations in terms of components and their behavior in terms of failure logic. The proposal is built on top of the tool-supported model-based failure logic analysis technique called CHESS-FLA. CHESS-FLA supports the analysis of the component-based system architectures to understand what can go wrong at a system level, by applying failure logic rules at a component level. CHESS-FLA addresses only hardware and software components and as such it is inadequate for the analysis of socio-technical systems. This thesis proposes an extension of CHESS-FLA based on the preexisting classification (developed within SERA), of failures of socio entities. This extension combines CHESS-FLA and SERA - classification and delivers an approach named Concerto-FLA. Concerto-FLA is fully integrated into the CONCERTO framework allowing an automated analysis to be performed on architectures that contain human, organizational and technical entities present in socio-technical systems. The use of the approach is demonstrated on a case study extracted from the petroleum domain. The effectiveness of the delivered tool is briefly evaluated based on the results from the case study. / CONCERTO project

Data that can be represented as a network, where there are measurements both on units and on pairs of units, are becoming increasingly prevalent in the social sciences and public health. Homophily in network data, or the tendency of units to connect based on similar nodal attribute values (i.e. income, HIV status) more often than expected by chance is receiving strong attention from researchers in statistics, medicine, sociology, public health and others. Respondent-Driven Sampling (RDS) is a link-tracing network sampling strategy heavily used in public health worldwide that is cost efficient and allows us to survey populations inaccessible by conventional techniques. Via extensive simulation we study the performance of existing methods of estimating population averages, and show that they have poor performance if there is homophily on the quantity surveyed. We propose the first model-based approach for this setting and show its superiority as a point estimator and in terms of uncertainty intervals coverage rates, and demonstrate its application to a real life RDS-based survey. We study how the strength of homophily effects can be estimated and compared across networks and different binary attributes under several network sampling schemes. We give a proof that homophily can be effectively estimated under RDS and propose a new homophily index. This work moves towards a deeper understanding of network structure as a function of nodal attributes and network sampling under homophily. / Statistics

Diagnosis is the task of finding faults or malfunctioning components in a technical system, e.g a car. When doing diagnosis on cars with combustion engines, a computer program can be used. The computer program, also called diagnosis system, needs information about the car. This information could be data sheets of all the electronic components in the car. It could also be a description of how the engine behaves in a nominal and a non-nominal case. This information is contained in a model of the engine. RODON, a diagnostic tool developed by Sörman Information and Media AB, uses models of systems for conflict detection diagnosis. RODON needs fault models of the components to do diagnosis. The diagnosis system is then used in workshops, factories, or other places where cars need to be surveyed. In this thesis, a Simulink model of the nominal behaviour of a combustion engine is given. The problem is how to make use of the model as well as the diagnostic tool RODON for combustion engine diagnosis. To solve this, the Simulink model is translated into a RODON model. Translating a Simulink model into a RODON model requires a new library in RODON. The library developed in this thesis is called AdvancedBlocks library. The Simulink model describes the nominal behaviour of a combustion engine but for diagnosis with RODON, fault models are needed as well. Several types of faults that can occur in an engine have been studied and fault models have been implemented in RODON. The conclusion is that diagnosis in RODON with a translated engine model is possible.

We introduce a dimension reduction method for model-based clustering obtained from a finite mixture of t-distributions. This approach is based on existing work on reducing dimensionality in the case of finite Gaussian mixtures. The method relies on identifying a reduced subspace of the data by considering how much group means and group covariances vary. This subspace contains linear combinations of the original data, which are ordered by importance via the associated eigenvalues. Observations can be projected onto the subspace and the resulting set of variables captures most of the clustering structure available in the data. The approach is illustrated using simulated and real data. / Paul McNicholas

Model-based clustering allows for the identification of subgroups in a data set through the use of finite mixture models. When applied to high-dimensional microarray data, we can discover groups of genes characterized by their gene expression profiles. In this thesis, a mixture of skew-t factor analyzers is introduced for the clustering of high-dimensional data. Notably, we make use of a version of the skew-t distribution which has not previously appeared in mixture-modelling literature. Allowing a constraint on the factor loading matrix leads to two mixtures of skew-t factor analyzers models. These models are implemented using the alternating expectation-conditional maximization algorithm for parameter estimation with an Aitken's acceleration stopping criterion used to determine convergence. The Bayesian information criterion is used for model selection and the performance of each model is assessed using the adjusted Rand index. The models are applied to both real and simulated data, obtaining clustering results which are equivalent or superior to those of established clustering methods.