1234567 1 - 50 of 451
rss atomLink to result list
Permanent link
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
  • Möller, Clemens
    KTH, School of Architecture and the Built Environment (ABE), Sustainable development, Environmental science and Engineering, Industrial Ecology.
    Carbon Neutral Road Transportation: An Assessment of the Potential of Electrified Road Systems2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    Sweden is striving towards a carbon neutral transportation sector by2030 which includes reductions from CO2 emissions by 70%. This thesis focusses especially on the decarbonization of road freight transportation. Even though electrification of vehicles is seen as one of the available options to reach this goal, present battery technology does not meet requirementsof energy density and cost.

    The electrification of roads with electrified road systems (ERS) enables vehicles to charge electrical energy while in motion and has the potential toreduce weight and costs of on-board batteries for electric vehicles and avoids range anxiety of vehicle operators.

    Within this Master’s thesis, available ERSs are assessed and it is shown which of the available systems performs best in selected categories. Furthermore, alterantive options for large CO2 emission reductions in the roadtransportation sector are evaluated and it is shown that ERSs constitute the most promising alternative.

    Results of this dissertation are based on a qualitative research approachand limited to data availability.

  • Petrovic, Dario
    KTH, School of Architecture and the Built Environment (ABE), Real Estate and Construction Management.
    Risk Management in Construction Projects: A Knowledge Management Perspective from Swedish Contractors2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    Projects within the construction sector is characterized as fragmented, temporary and complex which inherently brings upon risk exposure. Decision makers within the industry need reliant access to information and knowledge in order to manage risks in a sufficient and systematic way. Thus, the implementation of an effective risk management in relation to managing associated project risk knowledge may facilitate successful construction project endeavors.  The purpose of this master thesis is to explore and evaluate project risk management within the Swedish construction industry, with the emphasis on the perspective of Swedish contractors. The aim is to examine the recognition and practical adoption of risk management in order to investigate how project knowledge is utilized in the process. The methodology consists of a literature review on risk management fundamentals, the diverse risk attitudes and knowledge management in relation to risk. Data collection and analysis are based on a mixed method approach in which conclusions are made in relation to theoretical groundwork. The empirical data is collected through the usage of an online survey and in-depth semi structured interviews with key professionals within the Swedish construction industry. The results from this research indicate that theoretical models and processes for risk management is fairly unknown within the industry, analogous methods are used within respective organizations but the sharing of definitions and concepts within the industry are absent and the methods are not as structured as described in risk management theory. However, findings suggest that the perception of actors within the industry regarding risk and the risk management process is one of high importance for obtaining project objectives. Furthermore, findings indicate that the interplay between knowledge- and risk management and the incorporation of these processes is underutilized among the contractors and developers.

  • Al-Nassrawi, Hamzah
    et al.
    KTH, School of Architecture and the Built Environment (ABE), Civil and Architectural Engineering, Concrete Structures.
    Tsamis, Grigorios
    KTH, School of Architecture and the Built Environment (ABE), Civil and Architectural Engineering, Concrete Structures.
    Global Analysis and design of a complex slanted High-Rise Building with Tube Mega Frame2017Independent thesis Advanced level (degree of Master (Two Years)), 80 credits / 120 HE creditsStudent thesis
    Abstract [en]

    The need for tall buildings will increase in the future and new building techniques will emerge to full fill that need. Tyrénshas developed a new structural system called Tube Mega Frame where the major loads are transferred to the ground through big columns located in the perimeter of the building. The new concept has the advantage of eliminating the core inside the heart of the building but furthermore gives countless possibilities and flexibility for a designer. The elimination of the central core, plus the multiformity the Tube Mega Frame, can result new building shapes if combined with new inventions like the Multi elevator Thussenkrupp developed.

    Multi is a new elevator system with the ability to move in all directions apart from vertically. In this thesis research of the possible combinations between TMF and Multi was conducted. The building shaped resulted is only one of the many possible outcomes which the mix of Multi and TMF can have. The building was constructed in a way so the TMF would be the main structural system, the building would have inclinations so the multi elevator would be the only elevator appropriate for the structure and the height would be significantly large.

    The pre-study focused on the inclination and its particularities. The inclination played a significant role on how the inner forces were distributed in a structure. Under special circumstances the inclination could be even beneficial although inclination could result in axial forces on the slabs so the horizontal elements should be designed thoroughly not only for bending or shear but also for axial loading. The next phase was experimenting on different simple shaped buildings and combinations of them. The conclusions on the simple buildings formed the idea on how the main building would be.

    The main building was modeled using four different structural systems and their subcategories with seven models in total. Totally seven systems were compared in load combinations for wind, dead, live, and seismic loads and the global behavior was studied. The model comparison included maximum deformations and modes of vibrations. This way the best structural systems were discovered for the specific building shape and conclusions on inclination into a structure were made. The best structural systems and more reliable in terms of results but also in simplicity of construction were chosen to be designed in ETABS. The 50m belt system, the outside braces system and the diagrid system were designed.

    The design of the buildings was conducted using the American code ASCE /SEI 7-10. In the design two different mega columns were used to study how a solid or hollow cross section can affect the global behavior. Depending on the structural system the mega column had a major or minor effect on the stiffness of the structure. The design of the cross sections was divided in many groups since the complex geometry had an impact on how and where forces arised in the structure. The outside brace system had the best results in terms of less weight and global stiffness proving that in inclined building and columns with the correct bracing and triangulation of elements could extinguish the negative effects of inclination and even perform better compared to conventional buildings.

     The 50-belt system was furthermore studied in buckling since it was one of the best structural systems but with the least bracing, but also the least complex in terms of construction method. The automated buckling through ETABS was conducted and a more conservative approach where the user is defining the buckling length and support factors was used. In addition, a comparison between the user defined factors and global buckling was conducted.

  • Kanjookaran, Noble Paul
    KTH, School of Industrial Engineering and Management (ITM), Applied Mechanical Engineering (KTH Södertälje).
    Identification and analysis of risks faced by Indian student community in Australia2017Independent thesis Advanced level (degree of Master (One Year)), 10 credits / 15 HE creditsStudent thesis
  • Karlsson, Rikard
    KTH, School of Electrical Engineering (EES), Electromagnetic Engineering.
    Power system performance when implementing dynamic rating on a wind farm connected transformer2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    In this study, dynamic transformer rating (DTR) is applied on a medium power transformer thatis in use in the regional transmission grid. The transformer’s rated apparent power is 63 MVA, itis connected between a wind farm at 22kV and the grid, K¨orsb¨arsdalen, at 135kV. Power systemanalysis are carried out on the grid, with the objective to test how DTR and increased wind poweraffects the grid performance with respect to reliability, voltage stability and active losses.Historical measurements of ambient temperature and transformer loading is used to calculatefree transformer capacity based on unity life time loss. For the investigated transformer and giventime period the load can be increased without endanger transformer lifetime. The dynamicallyrated capacity exceeds the nominal capacity through the whole tested time period but DTR showsmost effective during winter when the ambient temperature is colder and therefore, has a coolingeffect on the transformer. DTR is used to calculate available transformer capacity but an increasedcapacity often comes with the expense of decreased lifetime and reliability. With the help of DTRthe current rate of transformer lifetime usage can be calculated and used as input to economicalanalysis where there is a trade of between reliability and profit from increased capacity.Through the power system analysis procedure presented in this report it is possible to sortout the faults that are most probable to cause severe violations, this information can be used forefficient reinforcement of the grid. The contingency and reliability analysis can work as base tofind the solutions that will decrease the probability for violations most efficiently. Furthermore,the developed procedure can also be used when expanding the grid and to determine which gridalternative that will be most reliable.

  • Public defence: 2017-11-14 13:00 FB42, Stockholm
    Etcheverry Cabrera, Sebastian
    KTH, School of Engineering Sciences (SCI), Applied Physics, Laser Physics.
    Advanced all-fiber optofluidic devices2017Doctoral thesis, comprehensive summary (Other academic)
    Abstract [en]

    Significant technological advances of the last years have been possible by developments in Optofluidics, which is a field that deals with the integration of optics and microfluidics into single devices.

    The work described in this thesis is based on five scientific publications related to the use of fiber optic technology to build integrated optofluidic devices. The first three publications are within the field of life-science and point towards in-vivo and point-of-care applications, whereas the last two publications cover the study and the use of plasmonic nanoparticles for electrical modulation of light.

    Aiming at developing useful tools for in-vivo biological applications, the first publication consists of designing and testing a functional optical fiber for real-time monitoring and selective collection of fluorescent microparticles. This probe relies on a microstructured optical fiber with a hole along its cladding, which is used to selectively aspirate individual particles of interest once their fluorescence signal is detected. On the same line of research, the second publication contemplates the fabrication of a fiber probe that traps single microparticles and allows for remote detection of their optical properties. This probe is also based on a microstructured fiber that enables particle trapping by fluidic forces. The third publication addresses the development of an all-fiber miniaturized flow cytometer for point-of-care applications. This system can analyze, with excellent accuracy and sensitivity, up to 2500 cells per second by measuring their fluorescence and scattering signal. A novel microfluidic technique, called Elasto-inertial microfluidics, is employed for aligning the cells into a single-stream to optimize detection and throughput.

    The fourth publication involves the experimental and theoretical study of the electrical-induced alignment of plasmonic gold nanorods in suspension and its applicability to control light transmission. This study is done by using an all-fiber optofluidic device, based on a liquid-core fiber, which facilitates the interaction of light, electric fields, and liquid suspensions. Results show that nanorods can be aligned in microseconds, providing a much better performance than liquid-crystal devices. Finally, the fifth publication consists of an upgrade of the previous device by integrating four electrodes in the cladding of the liquid-core fiber. This improvement enables nanosecond response time and the possibility of digitally switching nanorods between two orthogonal aligned states, overcoming the limitation of slow thermal relaxation.

    The work presented here shows that optofluidics based on optical fibers is a robust and convenient platform, as well as a promising direction for the developing of novel instruments in fields such as life-science, non-linear optics, plasmonic, and sensing.

  • Thanopoulos, Sotirios
    KTH, School of Electrical Engineering (EES), Electromagnetic Engineering.
    Condition Monitoring of MV Remotely Controlled Distributed Disconnectors2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    During the last decades, the power grid is getting rapidly digitalised in order to contribute to theestablishment of Smart Grids and evaluate efficiently the extracted bidirectional data from the powersystem infrastructure. This thesis focuses on the MV grid, since its design and operation have changedmainly because of distributed generation installations and the increased demand of information fromstakeholders. Thus, asset management constitutes a significant tool that can increase the reliability ofthe MV network’s operation and its level of control. Studies have shown that a maintenance planbased on condition monitoring of power system apparatuses would be more effective compared to theimplemented time-based scheduled maintenance.This project focuses on MV remotely controlled disconnectors since studies have shown that theirnumber of failures is double compared to manually operated ones. Since maneuverability andsecondary function are the causes of a major failure with the highest occurrence rate, motor currentmonitoring is studied in this thesis. Some devices that have the capability to monitor disconnectors’motor current, are presented. Additionally, the obtained max motor current measurements areevaluated through a parametric and a non-parametric statistical test. The main challenge of this thesisis to show whether the behaviour of motor current can be an indicator regarding to the disconnector’scondition status.Hence, the impact of different factors on the behaviour of motor current is investigated. It is concludedthat disconnectors without a failure during the studied period are more likely to have max motorcurrent measurements higher than 8A and especially in the interval [10-12]A. The difference in motorcurrent of disconnectors with a work order and without failure is more significant in 2015/2016. Itseems that under the aforementioned values of max motor current, a disconnector is more probable tohave the capability to operate properly. It is also concluded that in case of malfunction “Mellanläge”,the value of max motor current is lower than 8A with higher probability and it maybe indicates aproblem of the studied disconnector.Through the comparison in pairs, it could be concluded that the effect of the external environmentalconditions is not so high on the behaviour of disconnectors’ max motor current measurements. Incontrast, it seems that the implementation of a work order, the number of operations and if adisconnector is installed more northerly in Zone 3 play a more significant role on the behaviour of thisdisconnector’s max motor current measurements. Consequently, based on the aforementioned results itis shown that some of the investigated factors could constitute an indicator whether a disconnector ismore or less probable to have the capability to operate properly.Finally, it is calculated the reduction in the interruption cost that could be achieved in case ofimplementation of motor current monitoring on Vattenfall’s remotely controlled distributeddisconnectors.

  • Svanberg, Rikard
    KTH, School of Industrial Engineering and Management (ITM), Materials Science and Engineering, Energy and Furnace Technology.
    Ex-situ Ion Enhanced Pyrolysis of Biomass: Effects of low power high voltage spark on the pyrolysis products2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
  • Hargedal, Axel
    et al.
    KTH, School of Architecture and the Built Environment (ABE), Real Estate and Construction Management.
    Danmo, Emil
    KTH, School of Architecture and the Built Environment (ABE), Real Estate and Construction Management.
    Bedömning av kommersiella fastighetskrediter: En studie om förhållandet mellan kreditgivarens bedömning avkassaflöde och värderingsflöde2017Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    The study has, using a qualitative research method through interviews investigated creditors assessment procedure for commercial real estate mortgages. An area of interest to study since the purpose of credit risk assessment is about evaluating the risk of the creditor not getting back the money that’s been lent. The purpose has been to investigate the overall procedure and the assessment of cash flow for both repayment ability and valuation of collateral purposes and their roles in the credit risk assessment process. The study investigates the different parts of the procedure with regards to the four questions. It appears that the debtor’s repayment ability is measured through a detailed cash flow projected over a relatively short horizon. Valuation of property collateral is done using a less detailed cash flow projected over a longer horizon. The study concludes that the short-term cash flow almost exclusively is the central part in every assessment and among other findings that ratios referring to cash flow, that’s historically been used in the assessment of operating companies, are now being used for commercial real estate mortgages. The market value of collateral has proven to be less significant even during times of economic crisis when financial covenants referring to these are often broken, but still accepted, if the repayment ability withstands. This seems natural since the prosperity of the engagement lies within the ability of the cash flow to cover the loan servicing, not within the value of the collateral. A claim and liquidation of collateral in the best of cases only results in repayment of the outstanding loan balance. A few cases where the value of collateral receives a greater importance have however during the study been identified.

  • Karaoguz, Hakan
    et al.
    KTH, School of Computer Science and Communication (CSC), Robotics, perception and learning, RPL.
    Bore, Nils
    KTH, School of Computer Science and Communication (CSC), Robotics, perception and learning, RPL. KTH, School of Computer Science and Communication (CSC), Centres, Centre for Autonomous Systems, CAS.
    Folkesson, John
    KTH, School of Computer Science and Communication (CSC), Robotics, perception and learning, RPL. KTH, School of Computer Science and Communication (CSC), Centres, Centre for Autonomous Systems, CAS.
    Jensfelt, Patric
    KTH, School of Computer Science and Communication (CSC), Robotics, perception and learning, RPL. KTH, School of Computer Science and Communication (CSC), Centres, Centre for Autonomous Systems, CAS.
    Human-Centric Partitioning of the Environment2017Conference paper (Refereed)
    Abstract [en]

    In this paper, we present an object based approach for human-centric partitioning of the environment. Our approach for determining the human-centric regionsis to detect the objects that are commonly associated withfrequent human presence. In order to detect these objects, we employ state of the art perception techniques. The detected objects are stored with their spatio-temporal information inthe robot’s memory to be later used for generating the regions.The advantages of our method is that it is autonomous, requires only a small set of perceptual data and does not even require people to be present while generating the regions.The generated regions are validated using a 1-month dataset collected in an indoor office environment. The experimental results show that although a small set of perceptual data isused, the regions are generated at densely occupied locations.

  • Public defence: 2017-11-13 09:00 Kollegiesalen, Stockholm
    Fridlund, Gustav
    KTH, School of Architecture and the Built Environment (ABE), Urban Planning and Environment, Urban and Regional Studies.
    Utan tvivel är en inte klok: En studie om personliga skavningar som resurs för praktisk klokhet inom svensk kommunal planering2017Doctoral thesis, monograph (Other academic)
    Abstract [en]

    How can you as a planner tackle messy realities without losing sight of possible problematic outcomes of what you put in practice? This study explores the value of everyday frictions as a resource for phronetic planning, i.e. the ability to make situated ethical judements of what is ’better’ or ’worse’ in a particular setting. The intent is to offer a situated gaze of frictions from the perspective of a civil servant of the well organised and innovative municipality of Botkyrka in the metropolitan area of Stockholm, Sweden. From this outset, an autoethnographical methodology from a poststructural approach, is used to explore the frictions that the author has experienced as a practicing planner. The study shows that frictions can be used as 'weak signals' to identify possible tricky consequences of the creation and the staging of planning 'simplifications'. Based on this insight a 'seismological' approach to planning is proposed. The argument is that planning practice should on the one hand utilize frictions when they arise and, on the other hand, actively challenge existing 'simplifications'. To achieve this, practical tools are offered to 'evoke', 'narrate' and 'diffract' on frictions and 'trickster-objects' within the constraints of the planner’s role as a municipal civil servant. 

  • Akhand, Victoria
    KTH, School of Chemical Science and Engineering (CHE).
    Mass Balance Model of Impurities for the WoodRoll Process2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    The depletion time for fossil fuels calculated from 2009 is 35 years for oil, 37 years for gas

    and 107 years for coal. Hence, it has become essential to reduce the dependence on fossil

    fuels by switching over to renewable resources. This in turn will also help in combating the

    negative effects on the environment like global warming. Thermochemical processes such as

    pyrolysis and gasification of biomass are considered the most efficient technology for

    converting biomass to useful energy carriers. Cortus Energy AB is a Swedish company that

    has developed a patented gasification technology called WoodRoll

    ® for gasification of fuels derived from biomass. However, ash in a thermochemical conversion process can cause

    corrosion, sintering, slag and poisoning of catalysts, which leads to operational problems. In

    addition, heavy metals such as Zn, Pb and Cd are environmentally problematic. These metals

    contribute to environmental pollution by contaminating the soil, which in turn can harm

    humans and the ecosystem via the food chain. The undesired elements should therefore be

    identified to minimize their impact on the overall thermochemical process and to reduce the

    emission of these harmful substances.

    The objective of this master thesis project, on behalf of Cortus Energy AB, was to investigate

    possible key numbers that can be used to describe and predict how the ash behaves in their

    patented WoodRoll® process. The key numbers that have been identified are empirically

    developed based on experience of coal combustion. These key numbers are regularly used for

    fuel derived from biomass by companies specializing in analyzing, although knowledge about

    whether they can actually be used on biomass is limited. In order to ensure that the use of

    these for biomass is correct, they must be experimentally verified in the future. In addition, a

    theoretical investigation is conducted to study which species can form and in what phase this

    occurs. The investigation reveals that there are no clear trends for how the inorganic elements

    behave since contradicting results from different studies have been reported. Formation of

    species and their phase depends on several parameters such as temperature, heating rate,

    particle size, volatility, quantity and interaction between the elements in the biomass. The

    thesis project ends with a mass balance model on selected inorganic elements for wood

    residues, as well as for bark. The model could be a tool for Cortus Energy AB to identify

    approximately how much of each element is present in each stream in the WoodRoll® process.

    The models are verified with analysis results. Inorganic elements that affect the

    overall process and its equipment have been selected for modelling. The volume percent of

    H2, CO, CO2 and CH4 in the models agree well with the values obtained by Cortus Energy

    AB. This shows that the thermodynamic equilibrium calculations performed are reasonable.

  • Karlsson, Elin
    KTH, School of Architecture and the Built Environment (ABE), Urban Planning and Environment.
    Sociala aspekter av tillgänglighet: En studie av forskningsprojektet Sociala planeringsunderlag2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
  • Letzner, Josefine
    KTH, School of Computer Science and Communication (CSC).
    Analysis of Emergency Medical Transport Datasets using Machine Learning2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    The selection of hospital once an ambulance has picked up its patient is today decided by the ambulance staff. This report describes a supervised machinelearning approach for predicting hospital selection. This is a multi-classclassification problem. The performance of random forest, logistic regression and neural network were compared to each other and to a baseline, namely the one rule-algorithm. The algorithms were applied to real world data from SOS-alarm, the company that operate Sweden’s emergency call services. Performance was measured with accuracy and f1-score. Random Forest got the best result followed by neural network. Logistic regression exhibited slightly inferior results but still performed far better than the baseline. The results point toward machine learning being a suitable method for learning the problem of hospital selection.

  • Bergman, Gustav
    et al.
    KTH, School of Computer Science and Communication (CSC).
    Norén, Felix
    KTH, School of Computer Science and Communication (CSC).
    Graphical elements that can invoke trust in online web shops2017Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    Online stores are becoming a natural part of our daily life, and more and more services and products are purchased through an online store. As e-commerce is growing, so is computer fraud, and trustworthiness is now an important topic in e-commerce. Exploring what components in an e-commerce experience, from a customer's point of view, that affect trust is therefore important. The web page can be equated to the look of a physical store, but a company needs to have other trust indicating elements in an online store since a consumer cannot touch the product or have an eye to eye dialog with the salesman. So, the question is How do you display things in an online store to appeal trust in the first impression of the shop and what other graphical factors matterswhen you want to appeal trust? Our definition of trust is taken from McKnight and Chervany and is defined as "the extent to which one party is willing to depend on the other party in a given situation with a feeling of relative security, even though negative consequences are possible". We also build our study on how different colors get people to react in a certain way and the importance of the first impression. To answer the question about how to appeal trust in an online shop, we sent out a form to students at KTH Royal Institute of Technology containing 31 images of online stores and asked them to answer the question "Does this web shop seem trustworthy to you?" on each of these images. When creating the various images with different designs we wanted to include designs that we thought were going to be perceived as trustworthy, and designs that we thought not. We could then change some aspects in those designs to see what works or not. The result shows that online stores that are deemed professional is the most trustworthy ones and we can conclude that intense colors are something to avoid, while having a certificate and providing contact information is something to recommend.

  • Rude, Hampus
    et al.
    KTH, School of Computer Science and Communication (CSC).
    Balicevac, Aleksandar
    KTH, School of Computer Science and Communication (CSC).
    Byggnadsinformationsmanagement inom fastighetsförvalltning: En undersökning av hur temperaturdata från branddetektorer kan användas för att beskriva inomhusklimat samt vilka effektiviseringar detta kan generera inom fastighetsförvaltning2017Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    Industry 4.0 is based on a basic process of digitizing information, linking information and devices, and automating processes. In the real estate industry, this process has led to the emergence of a new technology track - BIM - Building Information Management. This report investigates and gives a deeper explanation of BIM as a tool, method and mindset, both theoretically, by a background description and practical by a concrete example. The concrete example investigates how the temperature of a room can be described using modern fire detectors with built-in temperature gauge in real time. The result shows the correlation between temperature measured with a fire detector at ceiling height and temperature at work height set to 140cm from the floor. The result is then put in a broader perspective as a proposal for further investigation.

  • Görries, Simon
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
    ATHENA Space Telescope: Line of Sight Control with a Hexapod in the Loop2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    The Advanced Telescope for High-Energy Astrophysics, Athena, is an x-ray telescope with a 2000 kg mirror module mounted on a six degree of freedom hexapod mechanism. It is currently assessed in a phase A feasibility study as L-class mission in the European Space Agency’s Cosmic Vision 2015-2025 plan with a launch foreseen in 2028. The total mass of the spacecraft is approximately 8000 kg, which is mainly distributed to the mirror module on the one side and to the focal plane module on the other side of the telescope tube. Such a design is without precedent in any European mission and imposes several challenges on analysis and design of the complex line of sight pointing control system with the moving mirror module in the loop. Not only does the moving mirror mass lead to time-variant parameter uncertainties of the system (inertia), but also does the hexapod motion complicate the guidance algorithms and induce complex disturbances onto the SC attitude dynamics.

    These challenges have been approached in this thesis in three steps. First, a Matlab®/Simulink® library has been built up, including all components required to model a hexapod in pointing control simulations. This Hexapod Simulation Library includes a complete hexapod kinematic model, a simplified hexapod actuator model, a spacecraft attitude dynamic model with the hexapod in the loop, state determination algorithms with the hexapod in the loop as well as online guidance algorithms for combined spacecraft and hexapod maneuvers. Second, different operational scenarios have been designed and analyzed for comparison. Third, closed-loop pointing control simulations for a representative reference case study similar to the Athena spacecraft have been performed for first feasibility analysis and performance comparison of the different operational scenarios.

    With these simulations, first, the time-variant parameter uncertainties and complex disturbance torques caused by the moving mirror mass have been characterized. Thereby, the disturbance noise induced by hexapod actuator step quantization has been identified as a potential design driver that needs to be analyzed further in early phases of the project. Second, feasibility of a baseline pointing control concept has been shown, i.e. performing hexapod and spacecraft maneuvers sequentially in time. And third, possible improvements to the baseline concept have been analyzed providing up to 27% faster transition time between two observations for the simulated scenario by performing hexapod and spacecraft maneuvers simultaneously and applying a path optimized spacecraft trajectory.

  • Leborg, Sebastian
    KTH, School of Computer Science and Communication (CSC).
    Implementation and evaluation of selected Machine Learning algorithms on a resource constrained telecom hardware platform2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    The vast majority of computing hardware platforms available today are not desktop PCs. They are embedded systems, sensors and small specialized pieces of hardware present in almost every digital product available today. Due to the massive amount of information available through these devices we can find new and exciting ways to apply and benefit from machine learning. Many of these computing devices have specialized, resource-constrained architectures and it might be problematic to perform complicated computations. If such a system is under heavy load or has restricted performance, computational power is a valuable resource and costly algorithms must be avoided. \\This master thesis will present an in-depth study investigating the trade-offs between precision, latency and memory consumption of a selected set of machine learning algorithms implemented on a resource constrained multi-core telecom hardware platform. This report includes motivations for the selected algorithms, discusses the results of the algorithms execution on the hardware platform and offers conclusions relevant to further developments.

  • Magnusson, Wilhelm
    KTH, School of Computer Science and Communication (CSC).
    The EU General Data Protection Regulations and their consequences on computer system design2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    As of writing this thesis, the EU’s new data protection laws (GDPR) will start to apply within one year. The new regulations are poorly understood by many and rumours of varying accuracy are circling the IT industry. This thesis takes a look at the parts of the GDPR concerning system design and architecture, clarifying what they mean and their consequences for system design. The new regulations are compared to the old data protection laws (Directive 95/46/EC), showing how companies must alter their computer systems in order to adapt. Using evaluations of the old data protection laws predictions are made for how the GDPR will affect the IT industry going forward. One of the more important questions are what tools are available for companies when adapting to privacy protection regulations and threats. This thesis aims to identify the most common processes for this kind of system modification and compare their effectiveness in relation to the GDPR.

  • Bromark, Johanna
    et al.
    KTH, School of Computer Science and Communication (CSC).
    Hallman, Isabelle
    Uppspelningshastighet och inlärning: Hur uppspelningshastighet av en video påverkar studenters inlärning2017Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    Educational videos have become an increasingly popular complement to, or part of, university education. Earlier studies indicate positive effects of students using educational videos as part of their studies. Since the speech rate of a conversation is about 125-150 words per minute, it is slower than the speed in which the brain can absorb information. This study aims to examine whether it is possible to make studying through use of educational videos more efficient by increasing the playback speed. The overall research question is how the playback speed of an educational video influences the short-term learning of the presented material. An experimental study was conducted at three separate occasions spread out over two weeks and included 26 participants. Four groups of students saw the same video in different playback speeds: 1, 1.25, 1.5 and 2 times the original speed. The results indicate that the playback speed does not affect the short term learning of the presented information. This means that it is possible to digest the same information from a video in a shorter span of time by increasing the playback speed. It is therefore possible to make learning through educational videos more efficient by adjusting the playback speed. The results show that this is true for speeds up to double the original speed, corresponding to 320 words per minute.

  • Kindestam, Anton
    KTH, School of Computer Science and Communication (CSC).
    Graph-based features for machine learning driven code optimization2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    In this paper we present a method of using the Shortest-Path Graph Kernel, on graph-based features of computer programs, to train a Support Vector Regression model which predicts execution time speedup over baseline given an unseen program and a point in optimization space, based on a method proposed in Using Graph-Based Program Characterization for Predictive Modeling by Park et al.

    The optimization space is represented by command-line parameters to the polyhedral C-to-C compiler PoCC, and PolyBench is used to generate the data set of speedups over baseline.

    The model is found to produce results reasonable by some metrics, but due to the large error and the pseudo-random behaviour of the output the method, in its current form, must reluctantly be rejected.

  • Kück, Corvin
    KTH, School of Electrical Engineering (EES), Automatic Control.
    MPC Design For Autonomous Drifting2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    The goal of this thesis is to evaluate the performance of different controllers to keep a remotecontrolledvehicle in a sustained drift. A bicycle model and an empirical tyre model are used formodelling the vehicle. The parameters for the used Fiala tyre model are experimentally identifiedand the simulation results of the modelled vehicle are compared to measured experimental data. Itfollows a stability analysis of the modelled system. The system is then linearized around one ofthe drift equilibria to allow controller design. A state feedback controller is designed to stabilizethe system, the controller gains are optimized using a Linear Quadratic Regulator (LQR) design,subsequently a Model Predictive Controller (MPC) is designed. Finally, the performance of the 3controllers is evaluated for a simulation with a disturbance acting on the system.

  • Heien Bjonge, Ingrid
    et al.
    KTH, School of Industrial Engineering and Management (ITM), Applied Mechanical Engineering (KTH Södertälje).
    Hofflander, Marcus
    KTH, School of Industrial Engineering and Management (ITM), Applied Mechanical Engineering (KTH Södertälje).
    Automatiserad sortering avsekundär metallisk råvara2017Independent thesis Basic level (university diploma), 10 credits / 15 HE creditsStudent thesis
  • Melander, Anton
    et al.
    KTH, School of Computer Science and Communication (CSC).
    Halvord, Erik
    KTH, School of Computer Science and Communication (CSC).
    Blockchain - What it is, and a non-financial use case2017Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

     The interest in blockchain technology has risen dramatically in a short period of time. Conceptualized by the release of Bitcoin in 2008, the technology has grown at a steady pace, gaining traction over time with the interest of a large number o f industries.

    This paper explores the current state of the blockchain research space, providing a thorough description of the technology and its fundamental features. A literature study was conducted with the aim of defining a blockchain, where the result shows that two main characteristics are present; A blockchain is a data structure storing information in blocks where each block references the previous block. Further, data is distributed among blockchain network participants.

    Furthermore, this paper showcases the versatility of the technology by proposing a blockchain based model for digital voting at general meetings. While the model managed to provide a transparent means of tallying and verifying the outcome of a vote, it should be noted that a compromise had to be made in terms of trusting a third party.

  • Masko, David
    KTH, School of Computer Science and Communication (CSC).
    Calibration in Eye Tracking Using Transfer Learning2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    This thesis empirically studies transfer learning as a calibration framework for Convolutional Neural Network (CNN) based appearance-based gaze estimation models. A dataset of approximately 1,900,000 eyestripe images distributed over 1682 subjects is used to train and evaluate several gaze estimation models. Each model is initially trained on the training data resulting in generic gaze models. The models are subsequently calibrated for each test subject, using the subject's calibration data, by applying transfer learning through network fine-tuning on the final layers of the network. Transfer learning is observed to reduce the Euclidean distance error of the generic models within the range of 12-21%, which is in line with current state-of-the-art. The best performing calibrated model shows a mean error of 29.53mm and a median error of 22.77mm. However, calibrating heatmap output-based gaze estimation models decreases the performance over the generic models. It is concluded that transfer learning is a viable calibration framework for improving the performance of CNN-based appearance based gaze estimation models.

  • Renman, Casper
    KTH, School of Computer Science and Communication (CSC).
    Creating Human-like AI Movement in Games Using Imitation Learning2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    The way characters move and behave in computer and video games are important factors in their believability, which has an impact on the player’s experience. This project explores Imitation Learning using limited amounts of data as an approach to creating human-like AI behaviour in games, and through a user study investigates what factors determine if a character is human-like, when observed through the characters first-person perspective. The idea is to create or shape AI behaviour by recording one's own actions. The implemented framework uses a Nearest Neighbour algorithm with a KD-tree as the policy which maps a state to an action. Results showed that the chosen approach was able to create human-like AI behaviour while respecting the performance constraints of a modern 3D game.

  • Bojs, Robert
    et al.
    KTH, School of Computer Science and Communication (CSC).
    Feng, Benny
    KTH, School of Computer Science and Communication (CSC).
    Övervakad maskininlärning för att identifiera nya kunder på energimarknaden2017Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    This paper explores alternative ways for smaller actors on the energy market to identify potential customers using publicly available data and different machine learning algorithms. During recent years, price has been considered to have the biggest impact on the behaviour of the consumers on the energy market. Since the bigger actors on the market can use their economies of scale to lower their prices, smaller actors need to find alternative ways to reach out to consumers. The machine learning algorithms in this paper will use the sales data from a small energy company, operating in Sweden and attempt to find a connection between existing customers using their demographic properties. By acquiring a deeper knowledge of what differentiates consumers that are willing to purchase energy from the energy company and the other consumers, the energy company may increase their rate of successful sales. Due to the lack of customer data avilable coupled with a lack of relevant public data, the results in this paper are not conclusive. However, it provides a baseline for future research as the results may be more reliable when the number of customers purchasing energy from The Energy Company increases.

  • Madosh, Farzana
    et al.
    KTH, School of Computer Science and Communication (CSC).
    Lundsten, Erik
    KTH, School of Computer Science and Communication (CSC).
    Utvärdering av Amazon Machine Learning för taggsystem2017Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    How companies deal with machine learning is currently a highly-discussed topic, as it can facilitate corporate manual work by training computers to recognize patterns and thus automate the working procedure. However, this requires resources and knowledge in the field. As a result, various companies like Amazon and Google provide machine learning services without requiring the user to have deep knowledge in the area. This study evaluates Amazon Machine Learning program for a tag system with data from the media company Newstag. In order to make this evaluation, a larger amount of data with tags was obtained from the company. The result of the study indicates that Amazon's program do not work for multilabel classification.

  • Krinaki, Maria
    KTH, School of Computer Science and Communication (CSC).
    Usability of Content Management Systems on touchscreen mobile devices2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    Content Management Systems (CMS) can support the creation and maintenance of an e-commerce website and its content. Together with the advance of mobile technology and Internet, it is now possible for business owners to manage their online store using their mobile touchscreen devices. This thesis aims to define the usability problems of the existing CMSs when they are used on mobile touchscreen devices. Heuristic evaluation was performed in three CMSs in order to gain a better understanding of how CMS on touchscreen works and what are the major issues that could downgrade usability. Based on the results of the evaluation, a prototype was implemented and evaluated in order to find out if the design characteristics of the prototype can enhance usability. The impact of the prototype is discussed in terms of efficiency and heuristics violations. Moreover, design guidelines are suggested and uncovered usability problems for further research are provided.

  • Olofsson, Nina
    KTH, School of Computer Science and Communication (CSC).
    A Machine Learning Ensemble Approach to Churn Prediction: Developing and Comparing Local Explanation Models on Top of a Black-Box Classifier2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    Churn prediction methods are widely used in Customer Relationship Management and have proven to be valuable for retaining customers. To obtain a high predictive performance, recent studies rely on increasingly complex machine learning methods, such as ensemble or hybrid models. However, the more complex a model is, the more difficult it becomes to understand how decisions are actually made. Previous studies on machine learning interpretability have used a global perspective for understanding black-box models. This study explores the use of local explanation models for explaining the individual predictions of a Random Forest ensemble model. The churn prediction was studied on the users of Tink – a finance app.

    This thesis aims to take local explanations one step further by making comparisons between churn indicators of different user groups. Three sets of groups were created based on differences in three user features. The importance scores of all globally found churn indicators were then computed for each group with the help of local explanation models. The results showed that the groups did not have any significant differences regarding the globally most important churn indicators. Instead, differences were found for globally less important churn indicators, concerning the type of information that users stored in the app.

    In addition to comparing churn indicators between user groups, the result of this study was a well-performing Random Forest ensemble model with the ability of explaining the reason behind churn predictions for individual users. The model proved to be significantly better than a number of simpler models, with an average AUC of 0.93.

  • Shabo, Jack
    KTH, School of Computer Science and Communication (CSC).
    High Density Simulation of Crowds with Groups in Real-Time2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    To simulate crowds of people is of great social interest and is also believed to be useful when analyzing situations involving denser crowds. Many simulators seen over the years have however been struggling with simulating larger number of people, often due to a computationally expensive collision avoidance step. Furthermore, many simulators seems to forget the fact that people tend to stick together in smaller social groups rather than walking alone.

    A simulator for high density crowds has nevertheless been implemented through modeling crowds as a unilateral incompressible fluid. Together with an integration of groups onto this approach, the obtained solution allows for real time simulation of up to 3000 virtual people.

    The impact of having groups in simulations has furthermore been set as the overall goal of the thesis and has been analyzed through observing the effects of groups in various scenarios. A smaller user study has also been conducted in order to gain perceptual insights of groups in various crowd densities. These have shown that groups have a smaller impact on the crowd flow, and do not put a larger strain on the performance of the simulation. Groups are further proved to be perceived differently in different densities, with a possible difficulty for scenarios in higher density.

  • Dönmez, Halit Anil
    KTH, School of Computer Science and Communication (CSC).
    Collision Avoidance for Virtual Crowds Using Reinforcement Learning2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    Virtual crowd simulation is being used in a wide variety of applications such as video games, architectural designs and movies. It is important for creators to have a realistic crowd simulator that will be able to generate crowds that displays the behaviours needed. It is important to provide an easy to use tool for crowd generation which is fast and realistic. Reinforcement Learning was proposed for training an agent to display a certain behaviour. In this thesis, a Reinforcement Learning approach was implemented and the generated virtual crowds were evaluated. Q Learning method was selected as the Reinforcement Learning method. Two different versions of the Q Learning method was implemented. These different versions were evaluated with respect to state-of-the-art algorithms: Reciprocal Velocity Obstacles(RVO) and a copy-synthesis approach based on real-data. Evaluation of the crowds was done with a user study. Results from the user study showed that while Reinforcement Learning method is not perceived as real as the real crowds, it was perceived almost as realistic as the crowds generated with RVO. Another result was that, the perception of RVO changes with the changing environment. When only the paths were shown, RVO was perceived as being more natural than when the paths were shown in a setting in real world with pedestrians. It was concluded that using Q Learning for generating virtual crowds is a promising method and can be improved as a substitute for existing methods and in certain scenarios, Q Learning algorithm results with better collision avoidance and more realistic crowd simulation.

  • Efraimsson, Nils
    KTH, School of Computer Science and Communication (CSC), Speech, Music and Hearing, TMH.
    Onset detection in polyphonic music2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    In music analysis, the beginning of events in a music signal (i.e. sound onset detection) is important for such tasks as sound segmentation, beat recognition and automatic music transcription. The aim of the present work was to make an algorithm for sound onset detection with better performance than other state-of-the-art1 algorithms. Necessary theoretical background for spectral analysis on a sound signal is given with special focus on the Short-Time Fourier Transform (STFT) and the effects of applying a window to a signal. Previous works based on different approaches to sound onset detection were studied, and a possible improvement was observed for one such approach - namely the one developed by Bello, Duxbury, Davies, & Sandler (2004). The algorithm uses an STFT approach, analyzing a sound signal time frame by time frame. The algorithm’s detection is sequential in nature: It takes a frame from the STFT and makes an extrapolation to the next frame, assuming that the signal is constant. The difference between the extrapolated frame and the actual frame of the STFT constitutes the detection function. The proposed improvement lies in a combination of ideas from other algorithms, analyzing the signal with different frequency bands with frequency dependent settings and a modification of the extrapolation step. The proposed algorithm is compared to the original algorithm and an adaption by Dixon (2006) by analyzing 20 songs using three different window functions. The results were evaluated with the standards set by MIREX (2005-2016). The results of the proposed algorithm are encouraging, showing good recall, but fail to out-perform any of the algorithms it is compared to in both precision and the so-called F-measure. The shortcomings of the proposed algorithm leave room for further improvement, and a number of possible future modifications are exemplified.

  • Jonsson, Fredrik
    et al.
    KTH, School of Computer Science and Communication (CSC).
    Tornkvist, Martin
    KTH, School of Computer Science and Communication (CSC).
    RSA authentication in Internet of Things: Technical limitations and industry expectations2017Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    The objective of this thesis is to evaluate if it is possible to run RSA authentication in a specified scenario. A Raspberry Pi with a limited CPU is used to simulate a low-performance device. A series of tests on this device shows that it is not possible to run RSA authentication in the provided scenario.A survey conducted on IT-professionals shows that there is a strong belief that this is possible. The results shows that there is a disparity between the tested RSA performance and the perception in the industry. However since ambiguity exists in the scenario it is hard to draw conclusions about the results.

  • Svärd, Mikael
    et al.
    KTH, School of Computer Science and Communication (CSC).
    Rumman, Philip
    KTH, School of Computer Science and Communication (CSC).
    COMBATING DISINFORMATION: Detecting fake news with linguistic models and classification algorithms2017Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    The purpose of this study is to examine the possibility of accurately distinguishing fabricated news from authentic news stories using Naive Bayes classification algorithms. This involves a comparative study of two different machine learning classification algorithms. The work also contains an overview of how linguistic text analytics can be utilized in detection purposes and an attempt to extract interesting information was made using Word Frequencies. A discussion of how different actors and parties in businesses and governments are affected by and how they handle deception caused by fake news articles was also made. This study further tries to ascertain what collective steps could be made towards introducing a functioning solution to combat fake news. The result swere inconclusive and the simple Naive Bayes algorithms used did not yieldfully satisfactory results. Word frequencies alone did not give enough information for detection. They were however found to be potentially useful as part of a larger set of algorithms and strategies as part of a solution to handling of misinformation.

  • Okurowska, Magdalena
    et al.
    KTH, School of Computer Science and Communication (CSC).
    Granlund, Linnéa
    KTH, School of Computer Science and Communication (CSC).
    En studie i studenters tv-tittande: Aspekter av binge-watching och maratontittande2017Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    Since television was invented one has been limited to linear television, both in terms of programs and the number of episodes. The number of channels has since then increased as has the number of available programs. But the watching still takes place on the same TV-set. With the introduction of the Video on Demand services (VoD) one got the opportunity to watch TV series wherever and wherever one wants. This has triggered new viewing patterns. One is binge-watching, a concept that many are unfamiliar with, even though many actually watch TV in that way. The term means that the consumer watch several episodes of a TV series at a time. Most digital services try, due to financial reasons, to extend the time the viewers spend at the TV, and therefore they develop platforms with features that encourage binge-watching. It is in the interest of the whole community to understand why and how we are persuaded to look longer on series in order to understand our viewing behavior and thereby stay in control.

    The purpose of our study is to investigate what factors that encourage the most to binge-watching. To answer this we also choose to look into: "What factors in Netflix and Viaplay encourage binge-watching the most?" and "Which external factors encourage binge-watching?". To find answers we took help from 17 media technology students at KTH that daily during a week was asked to logg their viewing habits. We have also conducted semistructured interviews with them to get in-depth knowledge of their use of VoD-services. Our work provides interesting and useful results based on the qualitative and quantitative methods we used, and is a good foundation for future studies of television viewing aimed at better understand the binge-watching phenomena.

    Our study shows that the autoplay function encourage viewers to continue to watch series. Only five of our study participants could resist this feature. We also conclude that the comfort-level of the viewing environment and the choice of electronic viewing device influence the tendency to marathon view. More than half of our study participants prefer to watch series on a larger screen. Many of them acknowledged that a TV-sized screen and a comfortable couch or bed encouraged binge-watching. Not only technical features encourage binge-watching. Also the time the user can devote to their viewing and the reason for viewing influence it.

    The study showed that many features in both the studied VoD-services are very similar. The biggest difference between them is that it in Viaplay is possible to pause the autoplay function while you can not do this in Netflix. But none of our participants knew that Viaplay had this feature and we could therefore not draw any conclusions as to whether this difference affected the binge-watching or not.

  • Jovanovic, Tamara
    KTH, School of Industrial Engineering and Management (ITM), Applied Mechanical Engineering (KTH Södertälje).
    Produktionsplaneringinom en servicemyndighet2017Independent thesis Basic level (university diploma), 10 credits / 15 HE creditsStudent thesis
  • Public defence: 2017-11-03 13:00 F3, Stockholm
    Karagulyan, Davit
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.).
    Certain results on the Möbius disjointness conjecture2017Doctoral thesis, comprehensive summary (Other academic)
    Abstract [en]

    We study certain aspects of the Möbius randomness principle and more specifically the Möbius disjointness conjecture of P. Sarnak. In paper A we establish this conjecture for all orientation preserving circle homeomorphisms and continuous interval maps of zero entropy. In paper B we show, that for all subshifts of finite type with positive topological entropy the Möbius disjointness does not hold. In paper C we study a class of three-interval exchange maps arising from a paper of Bourgain and estimate its Hausdorff dimension. In paper D we consider the Chowla and Sarnak conjectures and the Riemann hypothesis for abstract sequences and study their relationship.

  • Rahme, Daniel
    KTH, School of Technology and Health (STH), Medical Engineering, Computer and Electronic Engineering.
    Utveckling och mätning av följdetekteringsalgoritm för V2V med hänsyn till tidsfördröjning2017Independent thesis Basic level (university diploma), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    Traffic accidents are a problem in today’s society and new solutions regarding intelligent transportation systems (ITS) are being requested. One solution to such is vehicle-to-vehicle communication (V2V), where safer traffic can be achieved by vehicles transmitting relevant traffic information to surrounding vehicles. In this report, we researched and developed methods for a follow-detection algorithm for V2V communication. The task was to find a solution for follow-detection by using GPS and FM-RDS, with regards to time delay. An algorithm was developed in a simulated environment, to then be implemented in a prototype for measuring traffic. For efficient usage of FM-RDS packets, the transmitted position was compressed and then decompressed at the receiving side. The results of the simulations and traffic tests indicated that it was possible to detect if a vehicle followed another in traffic, by using GPS positioning and FM-RDS with time delay. The accuracy of the algorithm degraded at higher time delays, however there is potential for improvements. Suggestions for improvements of the reconstruction function and the algorithm are discussed in this report.

  • Rogstam, Jörgen
    et al.
    Energi & Kylanalys.
    Mazzotti, Willem
    KTH, School of Industrial Engineering and Management (ITM), Energy Technology, Applied Thermodynamics and Refrigeration.
    Ice Rink Dehumidification Systems Energy Usage and Saving Measures2014Conference paper (Refereed)
    Abstract [en]

    Ice skating rinks are one of the largest energy consumers in the building sector. The dehumidification function is one of the dominating energy users but is vital for the climate control in the ice rink. High humidity level causes extra diffusion load on the ice which affects the ice surface structure and risk to form condensation in the building structure. This study focuses on mass transfer phenomena in ice rinks. Three elements of the mass transfer mechanism are investigated: the mass transfer on the ice surface; the infiltration rate; and the dehumidifier energy consumption. The energy usage related to dehumidification of an ice rink is concluded to be as high as 15% of the total energy usage, which annually may correspond up to 150 MWh. A moisture balance model including infiltrations, moisture sources, condensation and dehumidification is proposed. A possible saving measure is to replace the electricity used for regeneration with recovered heat from the refrigeration system. Results from a case study show an electrical energy saving potential of 40% using other sources than electricity to heat the regeneration air.

  • Karagulyan, Davit
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.).
    Hausdorff dimension of a class of three-interval exchange mapsManuscript (preprint) (Other academic)
    Abstract [en]

    In \cite{B} Bourgain proves that Sarnak's disjointness conjecture holds for a certain class of Three-interval exchange maps. In the present paper we slightly improve the Diophantine condition of Bourgain and estimate the constants in the proof. We further show, that the new parameter set has positive, but not full Hausdorff dimension. This, in particular, implies that the Lebesgue measure of this set is zero.

  • Karagulyan, Davit
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.).
    On certain aspects of the Möbius randomness principleManuscript (preprint) (Other academic)
    Abstract [en]

    In this paper we study different aspects of the Möbius randomness principle. We rephrase the Chowla, Sarnak conjectures and the Riemann hypothesis for abstract sequences and study their relationships. We, in particular, show, that in this setting the Chowla and Sarnak conjectures do not imply the Riemann hypothesis. In the second part of the paper we also study the connection between the multiplicative and additive van der Corput criteria.

  • Public defence: 2017-11-06 10:00 T2, Huddinge
    S. Alvarez, Victor
    KTH, School of Technology and Health (STH), Medical Engineering, Neuronic Engineering.
    Understanding Boundary Conditions for Brain Injury Prediction: Finite Element Analysis of Vulnerable Road Users2017Doctoral thesis, comprehensive summary (Other academic)
    Abstract [en]

    Vulnerable road users (VRUs) are overrepresented in the statistics on severe and deadly injuries in traffic accidents, most commonly involving the head. The finite element (FE) method presents the possibility to model complex interactions between the human body and vehicles in order to better understand the injury mechanisms. While the rapid development of computer capacity has allowed for increasingly detailed FE-models, there is always a benefit of reducing the studied problem. Due to its material properties, the brain is more sensitive to rotational motion than to purely linear, resulting in complex injury causation. When studying brain injuries caused by a direct impact to the head, simulations using an isolated head model significantly increases efficiency compared to using a complete human body model. Also evaluation of head protective systems uses isolated mechanical head representations. It is not, however, established the extent to which the boundary conditions of the head determine the outcome of brain injuries.

    FE models of both the entire human body and the isolated head were used in this thesis to study the effect of the body, as well as active neck muscle tension, on brain injury outcome in VRU accidents. A pediatric neck model was also developed to enable the study of age-specific effects. A vehicle windscreen model was developed to evaluate the necessity of capturing the failure deformation during pedestrian head impacts.

    It was shown that the influence of the neck and body on brain injury prediction is greater in longer duration impacts, such as pedestrian head-to-windscreen impacts with an average difference of 21%. In accidents with shorter duration impacts, such as head-to-ground bicycle accidents, the average influence was between 3-12%. The influence did not consistently increase or limit the severity, and was dependent on the degree of rotation induced by the impact, as well as the mode of deformation induced in the neck. It was also shown that the predicted brain injury severity is dependent on capturing the large deformations of fractured windscreen, with the greatest effect near the windscreen frame. The pediatric neck model showed a large effect of age-dependent anatomical changes on inertial head loading, making it a promising tool to study the age-dependent effects in VRU accidents.

  • Gip Orreborn, Jakob
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Asset-Liability Management with in Life Insurance2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    In recent years, new regulations and stronger competition have further increased the importance of stochastic asset-liability management (ALM) models for life insurance firms. However, the often complex nature of life insurance contracts makes modeling to a challenging task, and insurance firms often struggle with models quickly becoming too complicated and inefficient. There is therefore an interest in investigating if, in fact, certain traits of financial ratios could be exposed through a more efficient model.

    In this thesis, a discrete time stochastic model framework, for the simulation of simplified balance sheets of life insurance products, is proposed. The model is based on a two-factor stochastic capital market model, supports the most important product characteristics, and incorporates a reserve-dependent bonus declaration. Furthermore, a first approach to endogenously model customer transitions is proposed, where realized policy returns are used for assigning transition probabilities.

    The model's sensitivity to different input parameters, and ability to capture the most important behaviour patterns, are demonstrated by the use of scenario and sensitivity analyses. Furthermore, based on the findings from these analyses, suggestions for improvements and further research are also presented.

  • Ahlin, Filip
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Internal model for spread risk under Solvency II2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    In May 2009 the European Commission decided on new regulations regarding solvency among insurance firms, the Solvency II Directive. The directive aims to strengthen the connection between the requirement of solvency and risks for insurance firms. The directive partly consists of a market risk module, in which a credit spread risk is a sub category.

    In this thesis a model for credit spread risk is implemented. The model is an extended version of the Jarrow, Lando and Turnbull model (A Markov Model for theTerm Structure of Credit Risk Spreads, 1997) as proposed by Dubrana (A Stochastic Model for Credit Spreads under a Risk-Neutral Framework through the use of an Extended Version of the Jarrow, Lando and Turnbull Model, 2011). The implementation includes the calibration of a stochastic credit risk driver as well as a simulation of bond returns with the allowance of credit transitions and defaults.

    The modeling will be made with the requirements of the Solvency II Directive in mind. Finally, the result will be compared with the Solvency II standard formula for the spread risk sub-module.

  • Starlander, Isak
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Counterparty Credit Risk on the Blockchain2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    Counterparty credit risk is present in trades offinancial obligations. This master thesis investigates the up and comingtechnology blockchain and how it could be used to mitigate counterparty creditrisk. The study intends to cover essentials of the mathematical model expectedloss, along with an introduction to the blockchain technology. After modellinga simple smart contract and using historical financial data, it was evidentthat there is a possible opportunity to reduce counterparty credit risk withthe use of blockchain. From the market study of this thesis, it is obvious thatthe current financial market needs more education about blockchain technology.

  • Hagsjö Renberg, Oscar
    et al.
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Hermansson, Oscar
    KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
    Large claims in non-life insurance2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    It is of outmost importance for an insurance company to apply a fair pricing policy. If the price is too high, valuable customers are lost to other insurance companies while if it’s too low – it nets a negative profit.

    To achieve a good pricing policy, information regarding claim size history for a given type of customer is required. A problem arises as large extremal events occur and affects the claim size data. These extremal events take shape in individually large claim sizes that by themselves can alter the distribution for what certain groups of individuals are expected to cost.

    A remedy for this is to apply what is called a large claim limit. Any claim exceeding this limit is thought of as being outside the scope of what is captured by the original distribution of the claim size. These exceeding claims are treated separately and have their cost distributed across all insurance takers, rather than just the group they belong to.

    So, where exactly do you draw this limit? Do you treat the entire claim size this way (exclusion) or just the bit that is exceeding the threshold (truncation)?

    These questions are treated and answered in this master’s thesis for Trygg-Hansa.

    For each product code, a limit was achieved in addition to which method for exceeding data that was best to use.

  • Tatsis, Nikolaos
    KTH, School of Information and Communication Technology (ICT).
    Constructions for efficient MDS diffusion layers2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    Matrices are widely used in Block Cipher Diffusion layers, usually chosen for offering maximal branch numbers, and allowing lightweight hardware implementations through their low XOR Count. When implemented in software however, is XOR Count the only metric?

    This project will utilize the parallelism provided by modern SIMD vector instructions to evaluate metrics, by implementing different algorithmic approaches. Timing their executions will hopefully provide some insight into the effect of the immense complexity of modern architectures and software tools on the expected outcome of an algorithm.

    A further focus will be the construction of matrices with interesting properties over large extension fields, which can apply in recent white-box cipher designs[BIT16]. This involves implementing a 1979 paper[Mac71], and the translation from mathematical paper to working implementation will involve many coding challenges.

    Both having an evaluation of matrix diffusion software implementation approaches and another method for finding matrices can ultimately help with the design of more efficient block ciphers.

  • Barai, Milad
    et al.
    KTH, School of Information and Communication Technology (ICT).
    Heikkinen, Anthony
    KTH, School of Information and Communication Technology (ICT).
    Impact of data augmentations when training the Inception model for image classification2017Independent thesis Basic level (professional degree), 10 credits / 15 HE creditsStudent thesis
    Abstract [en]

    Image classification is the process of identifying to which class a previously unobserved object belongs to. Classifying images is a commonly occurring task in companies. Currently many of these companies perform this classification manually. Automated classification however, has a lower expected accuracy. This thesis examines how automated classification could be improved by the addition of augmented data into the learning process of the classifier. We conduct a quantitative empirical study on the effects of two image augmentations, random horizontal/vertical flips and random rotations (<180◦). The data set that is used is from an auction house search engine under the commercial name of Barnebys. The data sets contain 700 000, 50 000 and 28 000 images with each set containing 28 classes. In this bachelor’s thesis, we re-trained a convolutional neural network model called the Inception-v3 model with the two larger data sets. The remaining set is used to get more class specific accuracies. In order to get a more accurate value of the effects we used a tenfold cross-validation method. Results of our quantitative study shows that the Inception-v3 model can reach a base line mean accuracy of 64.5% (700 000 data set) and a mean accuracy of 51.1% (50 000 data set). The overall accuracy decreased with augmentations on our data sets. However, our results display an increase in accuracy for some classes. The highest flat accuracy increase observed is in the class "Whine & Spirits" in the small data set where it went from 42.3% correctly classified images to 72.7% correctly classified images of the specific class.

  • Shahrokhian, Daniyal
    KTH, School of Information and Communication Technology (ICT).
    Syna: Emotion Recognition based on Spatio-Temporal Machine Learning2017Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
    Abstract [en]

    The analysis of emotions in humans is a field that has been studied for centuries. Through the last decade, multiple approaches towards automatic emotion recognition have been developed to tackle the task of making this analysis autonomous. More specifically, facial expressions in the form of Action Units have been considered until now the most efficient way to recognize emotions. In recent years, applying machine learning for this task has shown outstanding improvements in the accuracy of the solutions. Through this technique, the features can now be automatically learned from the training data, instead of relying on expert domain knowledge and hand-crafted rules. In this thesis, I present Syna and DeepSyna, two models capable of classifying emotional expressions by using both spatial and temporal features. The experimental results demonstrate the effectiveness of Syna in constrained environments, while there is still room for improvement in both constrained and in-the-wild settings. DeepSyna, while addressing this problem, on the other hand suffers from data scarcity and irrelevant transfer learning, which can be solved by future work.