1234567 1 - 50 of 541
Cite
Citation style
• apa
• harvard1
• ieee
• modern-language-association-8th-edition
• vancouver
• Other style
More styles
Language
• de-DE
• en-GB
• en-US
• fi-FI
• nn-NO
• nn-NB
• sv-SE
• Other locale
More languages
Output format
• html
• text
• asciidoc
• rtf
• KTH, School of Electrical Engineering and Computer Science (EECS).
Synthetic Meta-Learning:: Learning to learn real-world tasks with synthetic data2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Meta-learning is an approach to machine learning that teaches models how to learn new tasks with only a handful of examples. However, meta-learning requires a large labeled dataset during its initial meta-learning phase, which restricts what domains meta-learning can be used in. This thesis investigates if this labeled dataset can be replaced with a synthetic dataset without a loss in performance. The approach has been tested on the task of military vehicle classification. The results show that for few-shot classification tasks, models trained with synthetic data can come close to the performance of models trained with real-world data. The results also show that adjustments to the data-generation process, such as light randomization, can have a significant effect on performance, suggesting that fine-tuning to the generation process could further improve performance.

• KTH, School of Electrical Engineering and Computer Science (EECS).
Machine Learning for Adaptive Cruise Control Target Selection2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Vehicles will be more complex, safe, and intelligent in the future. For instance, with the support of the advanced driver assistance system (ADAS), the safety and comfort of the driver and the passengers can be significantly improved. This degree project proposes data-driven solutions for adaptive cruise control (ACC) target selection that can be used to select one of the preceding vehicles as the primary target that similar to the choice of human drivers. This master degree project was carried out at Scania CV AB. A shared-network and a shared-LSTM network were used to select the primary target. Besides, A novel machine learning based target selection model (compare-target model) was designed, which can consider all neighboring vehicles together by comparing vehicles. A compare-target network and a compare-target XGBoost are developed based on the compare-target model. In total, four different machine learning methods were adopted to select the primary target for ACC, includ- ing a shared network, a shared-LSTM network, a compare-target network, and a compare-target XGBoost model. These methods were compared and analyzed. Fine-tuning was adopted to overcome the data imbalance problem of rare situations. The compare-target XGBoost can achieve 94.85% accuracy on the test set.

• KTH, School of Electrical Engineering and Computer Science (EECS).
Data augmentation using military simulators in deep learning object detection applications2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

While deep learning solutions have made great progress in recent years, the requirement of large labeled datasets still limit their practical use in certain areas. This problem is especially acute for solutions in domains where even unlabeled data is a limited resource, such as the military domain. Synthetic data, or artificially generated data, has recently attracted attention as a potential solution for this problem.

This thesis explores the possibility of using synthetic data in order to improve the performance of a neural network aimed at detecting and localizing firearms in images. To generate the synthetic data the military simulator VBS3 is used. By utilizing a Faster R-CNN architecture multiple models were trained on a range of different datasets consisting of varying amounts of real and synthetic data. Moreover, the synthetic datasets were generated following two different philosophies. One dataset strives for realism while the other foregoes realism in favor of greater variation. It was shown that the synthetic dataset striving for variation gave increased performance in the task of object detection when used in conjunction with real data. The dataset striving for realism gave mixed results.

• KTH, School of Electrical Engineering and Computer Science (EECS).
The digital transformation of the music industry through applications of blockchain technology2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Since its introduction in 2008, the blockchain technology has been hailed as the one that could revolutionise many different industries. The music industry underwent several changes in the last 20 years as a consequence of phenomena such as music piracy, digital music and music streaming. Many considered the blockchain technology the solution to the many issues the music industry is facing. However, while the technology has been around for more than 10 years, and despite the enthusiasm of scholars and experts, little has been done to actually implement blockchains in the industry, especially by its biggest players. The resulting question is if and how the technology could change the music industry. This research starts by looking into previous studies on the blockchain technology, on the music industry and on the possible intersections of the two: from there the most popular suggested applications of the blockchain technology are extracted and presented in interviews to people working in the music industry. The goal of the interviews is to understand if the issues that the suggestions aimed to solve are real, and if the suggested applications are thought to be actually useful. The results are also compared with considerations from previously analysed papers, and showed a general interest in the blockchain technology and in the belief that it could help solve some issues of the music industry. At the same time, it was found that the technology is considered too young to be employed at the present time, with most people reckoning it is a technology that could have an impact in 10-15 year time.

• KTH, School of Electrical Engineering and Computer Science (EECS).
Nonparametric Bayesian models for security anomaly detection2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Avionics systems are growing ever more complex to accommodate to the evolving needs of airlines. As such, their attack surface has greatly expanded, and the potential impact of a security incident touching these systems has become a critical concern for the aerospace industry. Countermeasures include defining security perimeters and monitoring network traffic, and in turn inspecting the logs generated by these systems to reveal security incidents and respond quickly accordingly. However, processing the large amount of system logs to extract valuable information is intractable through conventional means such as manual human investigation or regular expression matching.

This thesis tackles the problem of automation of anomaly detection on security and functional system logs using advanced machine learning techniques. It investigates new methods to improve on the work done during a previous project at Collins Aerospace on Markov chains and LSTM neural networks. In particular, it evaluates the use of nonparametric Bayesian methods to perform this task, specifically the Hidden Dirichlet Process Hidden Markov Model. A complete log analysis system is proposed based on these models, and their performance is evaluated on real-life datasets using this framework.

• KTH, School of Electrical Engineering and Computer Science (EECS).
Nonlinear Approximative Explicit Model Predictive Control Through Neural Networks: Characterizing Architectures and Training Behavior2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Model predictive control (MPC) is a paradigm within automatic control notable for its ability to handle constraints. This ability come at the cost of high computational demand, which until recently has limited use of MPC to slow systems. Recent advances have however enabled MPC to be used in embedded applications, where its ability to handle constraints can be leveraged to reduce wear, increase efficiency and improve overall performance in everything from cars to wind turbines. MPC controllers can be made even faster by precomputing the resulting policy and storing it in a lookup table. A method known as explicit MPC.

An alternative way of leveraging precomputation is to train a neural network to approximate the policy. This is an attractive proposal both due to neural networks ability to imitate policies for nonlinear systems, and results that indicate that neural networks can efficiently represent explicit MPC policies. Limited work has been done in this area. How the networks are setup and trained therefore tends to reflect recent trends in other application areas rather than being based on what is known to work well for approximating MPC policies. This thesis attempts to alleviate this situation by evaluating how some common neural network architectures and training methods performs when used for this purpose. The evaluations are carried out through a literature study and by training several networks with different architectures to replicate the policy of a nonlinear MPC controller tasked with stabilizing an inverted pendulum.

The results suggest that ReLU activation functions give better performance than hyperbolic tangent and SELU functions; and that dropout and batch normalization degrades the ability to approximate policies; and that depth significantly increases the performance. However, the neural network controllers do occasionally exhibit problematic behaviors, such as steady state errors and oscillating control signals close to constraints.

• KTH, School of Electrical Engineering and Computer Science (EECS).
MahlerNet: Unbounded Orchestral Music with Neural Networks2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Modelling music with mathematical and statistical methods in general, and with neural networks in particular, has a long history and has been well explored in the last decades. Exactly when the first attempt at strictly systematic music took place is hard to say; some would say in the days of Mozart, others would say even earlier, but it is safe to say that the field of algorithmic composition has a long history. Even though composers have always had structure and rules as part of the writing process, implicitly or explicitly, following rules at a stricter level was well investigated in the middle of the 20th century at which point also the first music writing computer program based on mathematics was implemented.

This work in computer science focuses on the history of musical composition with computers, also known as algorithmic composition, using machine learning and neural networks and consists of two parts: a literature survey covering in-depth the last decades in the field from which is drawn inspiration and experience to construct MahlerNet, a neural network based on the previous architectures MusicVAE, BALSTM, PerformanceRNN and BachProp, capable of modelling polyphonic symbolic music with up to 23 instruments. MahlerNet is a new architecture that uses a custom preprocessor with musical heuristics to normalize and filter the input and output files in MIDI format into a data representation that it uses for processing. MahlerNet, and its preprocessor, was written altogether for this project and produces music that clearly shows musical characteristics reminiscent of the data it was trained on, with some long-term structure, albeit not in the form of motives and themes.

• KTH, School of Electrical Engineering and Computer Science (EECS).
Development of an Intelligent Embedded Interface for Interpreting Biosignals Recorded by Novel Wearable Devices2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

In recent years there has been a considerable development in the realm of sensing technologies, embedded systems, wireless communication technologies, nanotechnologies, and miniaturization has made it possible to create smart wearable systems that can record data from the human bodies and monitor our daily activities. Most expectations for the successful deployment of wearable devices and their tangible impact for the society is in healthcare. Nevertheless, its use has been limited by the absence of intelligent mobile device interfaces able to process, analyse and inference the recorded data, giving relevant information to the user. On the other hand, new advances in nanotechnology have allowed the creation of so-called electronic skin, which consists in thin and flexible electrodes, easy and comfortable to use. This allows building new wearable devices able to record electrical activity from the surface of the body, which has a large diagnostic and monitoring potential.

In this work, the goal is to study the feasibility of using these new sensors for continuous biopotential recording while supporting them with a mobile phone application able to receive, process and analyse the recorded biosignals in order to deliver useful feedback to the user in real-time. The wearable device known as Senso Medical Bio Pot V2 is proposed as a possible candidate to carry out electromyography (EMG), electrocardiography (ECG) and electroencephalography (EEG) recordings via skin tattoo electrodes. Moreover, an Android Application that connects to this device is created. It uses Machine Learning Algorithms embedded on it in order to classify the received signals. Finally, Long-Short Term Memory (LSTM) networks are implemented for classifying EEG and EMG signals.

Several conclusions are derived from this work. Firstly, the device Senso Medical Bio Pot V2 is not suitable for its use as a wearable device for biosignal recording. Secondly, the Application designed and simulated offline achieves good performance. As a consequence, it could be used in the future with a suitable wearable sensor and offer good potential for processing and interpreting recorded biosignals with an opportunity to provide real-time feedback to the user in Brain-Computer Interface (BCI) type of applications.

• KTH, School of Electrical Engineering and Computer Science (EECS).
Interactive Visual Analytics for Agent-Based simulation: Street-Crossing Behavior at Signalized Pedestrian Crossing2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

To design a pedestrian crossing area reasonably can be a demanding task for traffic planners. There are several challenges, including determining the appropriate dimensions, and ensuring that pedestrians are exposed to the least risks. Pedestrian safety is especially obscure to analyze, given that many people in Stockholm cross the street illegally by running against the red light. To cope with these challenges, computational approaches of trajectory data visual analytics can be used to support the analytical reasoning process. However, it remains an unexplored field regarding how to visualize and communicate the street-crossing spatio-temporal data effectively. Moreover, the rendering also needs to deal with a growing data size for a more massive number of people.

This thesis proposes a web-based interactive visual analytics tool for pedestrians' street-crossing behavior under various flow rates. The visualization methodology is also presented, which is then evaluated to have achieved satisfying communication and rendering effectiveness for maximal 180 agents over 100 seconds. In terms of the visualization scenario, pedestrians either wait for the red light or cross the street illegally; all people can choose to stop by a buffer island before they finish crossing. The visualization enables the analysis under multiple flow rates for 1) pedestrian movement, 2) space utilization, 3) crossing frequency in time-series, and 4) illegal frequency. Additionally, to acquire the initial trajectory data, Optimal Reciprocal Collision Avoidance (ORCA) algorithm is engaged in the crowd simulation. Then different visualization techniques are utilized to comply with user demands, including map animation, data aggregation, and time-series graph.

• KTH, School of Industrial Engineering and Management (ITM).
Cartesian Force Estimation of a 6-DOF Parallel Haptic Device2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

The haptic device recreates the sense of touch by applying forces to the user. Since the device is “rendering” forces to emulate the physical interaction, the force control is essential for haptic devices. While a dedicated force/torque sensor can close the loop of force control, the additional equipment creates extra moving mass and inertia at the tool center point (TCP). Therefore, estimating the Cartesian force at the TCP has continuously been receiving attention over the past decades.

The objective of this thesis project is to develop a real-time force estimation algorithm based on the proportional current-torque relationship with the dynamic modeling of the TAU haptic device. The algorithm can be further used for the force control of the device. The research questions of the thesis are: how to design and develop an algorithm for the TAU that used for Cartesian contact force estimation, how to set up the force estimation test bench and how to evaluate the results of the force estimation algorithm.

In order to achieve the force estimation algorithm, a virtual environment is built to simulate the real-time haptic physics. Then an external force/torque sensor is installed at the TCP to get the measurement of the Cartesian force at the TCP. The force estimation algorithm calculates the Cartesian force at the TCP based on the current measurement of the DC motors at the six joints. The estimation result of the Cartesian force at the TCP is then compared with the force/torque sensor measurement to determine if the estimation algorithm is sufficiently accurate. The analysis of the estimation accuracy emphasizes the feasibility of Cartesian force estimation on the TAU haptic device.

• KTH, School of Electrical Engineering and Computer Science (EECS), Intelligent systems, Micro and Nanosystems.
KTH, School of Electrical Engineering and Computer Science (EECS), Intelligent systems, Micro and Nanosystems. KTH, School of Electrical Engineering and Computer Science (EECS), Intelligent systems, Micro and Nanosystems. KTH, School of Electrical Engineering and Computer Science (EECS), Intelligent systems, Micro and Nanosystems. KTH, School of Electrical Engineering and Computer Science (EECS), Intelligent systems, Micro and Nanosystems.
Toward Industrial Exploitation of THz Frequencies: Integration of SiGe MMICs in Silicon-Micromachined Waveguide Systems2019In: IEEE transactions on rehabilitation engineering, ISSN 1063-6528, E-ISSN 1558-0024, Vol. 9, no 6, p. 624-636Article in journal (Refereed)

A new integration concept for terahertz (THz) systems is presented in this article, wherein patterned silicon-on-insulator wafers form all DC, IF, and RF networks in a homogeneous medium, in contrast to existing solutions. Using this concept, silicon-micromachined waveguides are combined with silicon germanium (SiGe) monolithic microwave integrated circuits (MMICs) for the first time. All features of the integration platform lie in the waveguide’s H-plane. Heterogeneous integration of SiGe chips is achieved using a novel in-line H-plane transition. As an initial step toward complete systems, we outline the design, fabrication, and assembly of back-to-back transition structures, for use at D-band frequencies (110ï¿œ170 GHz). Special focus is given to the industrial compatibility of all components, fabrication, and assembly processes, with an eye on the future commercialization of THz systems. Prototype devices are assembled via two distinct processes, one of which utilizes semiautomated die-bonding tools. Positional and orientation tolerances for each process are quantified. An accuracy of $\pm \text3.5\; μ \textm$, $\pm \text1.5 °$ is achieved. Measured $S$-parameters for each device are presented. The insertion loss of a single-ended transition, largely due to MMIC substrate losses, is 4.2ï¿œ5.5 dB, with a bandwidth of 25 GHz (135ï¿œ160 GHz). Return loss is in excess of 5 dB. Measurements confirm the excellent repeatability of the fabrication and assembly processes and, thus, their suitability for use in high-volume applications. The proposed integration concept is highly scalable, permitting its usage far into the THz frequency spectrum. This article represents the first stage in the shift to highly compact, low-cost, volume-manufacturable THz waveguide systems.

• Public defence: 2020-01-31 14:00 K1, Stockholm
KTH, School of Engineering Sciences in Chemistry, Biotechnology and Health (CBH), Chemical Engineering, Chemical Engineering.
Water density impact on water flow and mass transport in rock fractures2020Doctoral thesis, comprehensive summary (Other academic)

One way of taking care of spent nuclear fuel is to place it in a geological repository. In Sweden, a three-barrier system is planned. The system is based on encapsulating the fuel in copper canisters. These are surrounded by bentonite clay and buried under 500 m of bedrock. As a part of the safety assessment, the Q-equivalent model is used to quantify the possible release of radioactive material. This model also describes the rate at which corrosive agents carried by seeping water in rock fractures can reach the canisters, which may affect the longevity of the canisters.

The aim of this thesis was originally to develop an experimental, phys- ical model to visualize and validate the Q-equivalent model. However, the overarching theme of this work has been to study the effect of minor density differences that might be overlooked in experiments, both concentration- dependent and density-difference induced by light absorption.

In the initial diffusion and flow-experiment and associated calculations and simulations, it was found that simple Q-equivalent can describe and quantify the mass transport in both parallel and variable aperture fractures. However, this is the case only if the density difference between seeping water and clay pore water is insignificant. It was found in experiments with dyes used to visualise the flow and diffusion patterns that even minimal density differences could significantly alter the flow pattern. Density differences can result from concentration gradients or be induced by light absorption. TheQ-equivalent model was extended to account for density-induced flow. The importance of density-induced flow due to concentration gradients at the setting of a long-term repository for nuclear waste was evaluated. It was found that concentration gradients are able to induce rapid vertical up- or downward flow. This could increase the overall mass transport of radioactive material up to the biosphere or carry it downward to larger depths.

• KTH, School of Industrial Engineering and Management (ITM), Learning.
Varierande hinder och svårfångade möjligheter för språkintroduktionselever: En vansklig resa i tiden2019Independent thesis Advanced level (professional degree), 10 credits / 15 HE creditsStudent thesis

This qualitative study covers the subject of mathematics. The methods used are student interviews, meetings and discussions with teachers as well as review of mathematical test scores and a broad, inductive, thematic analysis of the interview data.It describes students’ perceptions of obstacles and opportunities in their schooling in mathematics prior to arriving in Sweden and in their Swedish second language introductory program. The results also provide the students’ views on their possibilities to pass in mathematics. A discussion concludes critical progress factors that need to be addressed in order for the students to have realistic possibilities to pass in mathematics.Swedish, as a new academic language for the students, is one of the biggest challenges in their mathematics education. The study describes similarities, but most of all differences between their schooling before and after arrival in Sweden. Among those are differences between authoritarian and authoritative styles of education and communication as well as differences in socio-culturally based behaviours of which the latter can generate huge learning obstacles if not addressed. Some students have manged to adopt effective working methods while others show a need for bonding, support and assistance to cope in the Swedish school characterized by freedom with responsibility.In 2017, the Swedish school inspection authority conducted a quality inspection of 42 schools engaged in second language introduction programmes (Skolinspektionen, 2017). The outcome revealed shortcomings, i.e. the possibilities for students to reach their goals in those programs varied significantly. The critical progress factors of this study address some of the shortcomings highlighted in the quality inspection, thus emphasizing the relevance of this study.

• KTH, School of Architecture and the Built Environment (ABE), Urban Planning and Environment, Urban and Regional Studies.
Medborgardialogens makt och vanmakt: Fallstudie av Rosens röda matta sex år efter invigning2019Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis

• KTH, School of Architecture and the Built Environment (ABE), Urban Planning and Environment, Urban and Regional Studies.
The Organic Pattern of Space:: A Space Syntax Analysis of Natural Streets and Street Segments for Measuring Crime and Traffic Accidents2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

The natural streets model is a research prototype that has been shown to perform better than the conventional GIS-based streets segments for explaining traffic flow and human movement. However, given its experimental status, a gap in the literature was identified. Therefore, the aim of this thesis was to contribute to the literature by investigating the wider applications of natural streets and observe whether a city’s spatial configuration (or structure) is related to outcomes of human behaviour and activity. In this case, the two previously unstudied outcomes were chosen: crime and traffic accidents. Taking an exploratory approach, Stockholm was chosen as the case study. Using the space syntax methodology, the street segments and natural streets connectivity was used to analyse whether accessibility or ‘potential through movement’ is associated with crime and traffic accidents. Two study areas were generated: a primary study area consisting of six nested zones and a secondary study area with hot spots and cold spots for events of crime and traffic accidents. To observe the statistical association between connectivity and events of crime and traffic accidents for natural streets and street segments, a classical regression model was used. The regression analysis showed that natural streets perform significantly better than street segments as they are better able to explain events of crime and traffic accidents. However, more so for traffic accidents. Most importantly, the topological structure or scaling characteristics of natural streets served as a better indicator for measuring human phenomena. The implication of this is that it could potentially be used to further the understanding of human activities in the context of the urban environment.

• KTH, School of Architecture and the Built Environment (ABE), Urban Planning and Environment, Urban and Regional Studies.
Retrofitting Accessibility in a Rapidly Expanding City: The Case of Bus Rapid Transit and Transit-Oriented Development in Dar es Salaam2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Urbanisation is a global trend, but in recent decades it has been occurring at particularly high rates in the Global South. Cities in Sub-Saharan Africa are facing a number of challenges as their populations grow, and among these, urban accessibility stands out as one of the most difficult to contend with. Meeting this challenge will require new solutions, and recently Bus Rapid Transit (BRT) and Transit-Oriented Development (TOD) have emerged as two potentially revolutionary innovations, especially when combined. In 2016, Dar es Salaam launched a BRT system, and in 2017 it introduced a TOD strategy in order to combat the city’s urban accessibility crisis. This study investigates the urban form at and around BRT stations in order to characterise BRT-TOD in the city and extract lessons for future implementations of BRT-TOD in Dar es Salaam and beyond. The investigation was carried out within the framework of urban morphology. It was found that at locations in the city centre, BRT has generally been able to integrate seamlessly into the built environment and achieve many of the core principles of TOD. In the rest of the city, however, BRT acts simply as a form of transit, with stations generally removed from the urban fabric and the local place. Some general issues are the large building setback, the number of informal vendors, private transport operators, exclusive new developments, lack of public spaces, the low quality of the pedestrian environment and the reinforcement of unsustainable forms of urban sprawl. This study aims to contribute to a growing discussion of sustainable urban accessibility solutions in the rapidly expanding cities of the Global South.

• KTH, School of Architecture and the Built Environment (ABE), Urban Planning and Environment, Urban and Regional Studies.
Tensions and Synergies Between Tactical Urbanism and Social Sustainability: A Case Study of the Sunset Triangle Plaza2018Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

For the past several years, the term social sustainability has gained a strong foothold within urban studies and has become a pervasive and trendy term that seems to be on everyone’s lips. Public space is widely acknowledged as an important urban feature, often in association with the social sustainability. As cities around the world are experiencing rapid population growth, creating meaningful and enjoyable public spaces is more important than ever.

The heightened interest in social sustainability, public spaces and placemaking (as a physical manifestation of social sustainability) has led to the emergence of several urban intervention movements, such as Tactical Urbanism. In 2012, for the first time in Los Angeles’ history, this tactic was used to transform a car trafficked street in Silver Lake into a pedestrian friendly public space: the Sunset Triangle Plaza. The aim of this thesis is to, by studying the use and function of the plaza after the conversion, highlight how a broad concept such as social sustainability can be understood from a relatively small-scale public space intervention.

The case study was conducted during the spring and summer of 2018, using a variety of data sources including interviews and observations of the plaza during February and March 2018. Two interviews were conducted with managers of the businesses directly adjacent to the plaza. Moreover, street surveys were conducted on two different occasions to ask the public about their use and opinions about the plaza. The results from the case study are presented to illustrate the real-life experience of the theories about social sustainability, public space and a discussion regarding “Whose Public Space?”, when applied at a local context. The findings were then further divided into three categories: usage (what type of activities did the installation enable?), users (for whom were they enabled?) and change (indicators of how the site has changed), reflecting the notions of Tactical Urbanism.

While certain changes have been merely “tactical”, others were more substantial; businesses flourished, traffic safety increased, the space has become a meeting place and therefore, it has now got an identity. Immediate change was evident in the process of the physical change when the plaza was constructed, but what has also followed is a continuous change. Even though the plaza with its painted dots may not look like much, a new space for engagement and interaction has been created, both physically and mentally. In addition, converting a street for the cars into a plaza dedicated to pedestrians is especially symbolic in a city like Los Angeles, where the automobile has been the predominant mean of transport for the last 60 years and instrumental in shaping the city’s layout. However, the case study also showed that it is one thing to launch a Tactical Urbanism initiative and another thing to maintain it and achieve long-term social changes.

The examined concepts and models to evaluate whether a public space can be considered successful are not always useful. The Sunset Triangle Plaza has certainly changed, but it has implied a continuous change – for better and for worse. Thus, this study also shows that it is evident that the idea of the “organically emerged” city can imply both opportunities and limitations.

• KTH, School of Architecture and the Built Environment (ABE), Urban Planning and Environment, Urban and Regional Studies.
Manifestation of Urban Segregation in the Urban Form2019Independent thesis Advanced level (degree of Master (One Year)), 10 credits / 15 HE creditsStudent thesis

Segregation in urban areas is a universal phenomenon. A combination of factors include but are not limited to: city form, planning policies, policies relating to the settlement of immigrants, economic policies, building of infrastructure and chronological events in the growth of a city. Through this thesis work, I would like to examine how urban form is different in areas of the city where racial/ economic segregation is prevalent in Stockholm city. The hypothesis which I would like to examine through study is : The development and maintenance of urban form(which stands to denote all elements relating to the urban area: roads, pavements, lighting fixtures, buildings, public services etc) is influenced by segregation in the city, and it reflects and in turn reinforces the prevalent segregation.

• KTH, School of Architecture and the Built Environment (ABE), Urban Planning and Environment, Geoinformatics.
Custom Base Maps for Utility Network Applications2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Web maps are today used more than ever. These are often displayed in geographical information system solutions. Web maps are often constructed in the terms of different layers where the bottom-most layer is called base map or background map. A specific kind of base maps are those that are the canvas for utility networks. These base maps can look vastly different from each other. Sometimes the base map can be overlooked. Currently there are different theories and opinions on how these base maps shall be designed and what information they shall contain. This thesis investigates a suitable base map design for utility network management and also creates some prototypes as a proof of concept. The thesis focus mostly on three factors, colour, information visualisation and symbology. This is investigated by using a user-centred design approach and comparing it to existing findings in literature and among map theories. The user-centred design process involves a case study performed with participants that work with utility networks in web map solutions. The research method is an iterative process where the participants are answering three quantitative surveys. The surveys includes prototypes which is refined for each step after analysing the participants answers. The thesis concludes that low saturated colours is good practise to enhance the network, this is both evident in literature as well as in the case study. Amount of lightness in the base map seem to be more of a opinionated matter and can highly differentiate depending on users device. Using hue is recommended to make borders between map elements more distinct. The base map shall only obtain the most relevant information and more specific information is recommended to be added through additional layers. Further studies need to be done in order to investigate the networks design and there is also a need to investigate how a map solution with several layers shall be design and how the relationships between those layers shall be constructed for utility network management

• KTH, School of Industrial Engineering and Management (ITM), Materials Science and Engineering.
Aging of FeCrAl Surface Coatings2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

This Thesis is about the aging of FECRAL surface coatings. In this thesis various substrates have been investigated such as 16 Mo3, 304, 347, Sanicro 31, 800HT, Nikrothal 80 and Kanthal APMT. These substrates have been coated with different FeCrAl alloys, using two different coating methods spraying with (High Velocity Air Force) and welding with (Metal Inert Gas).

The purpose of coating is to achieve specific properties of the layer without affecting the original properties of the substrate. Therefore, it is important to investigate the boundary layer between these two different materials to observe how the interdiffusion of different substances such as Cr, Al, Fe and C is affected, which is the purpose of this project.

The method used to investigate this purpose was to expose these combinations in different temperatures and in different environments for different time intervals. Then, using LOM, SEM and EDS analysis, the change that the boundary layers have undergone is examined. Some calculations in DICTRA have also been performed to see if it was possible to find any connection between experimental data and simulation results.

The result showed carburization of FeCrAl-coatings on 16Mo3 substrates which can lead to deterioration of mechanical properties in the substrates but also decreased corrosion resistance for the coated layers. The result has also shown that it is difficult to perform spraying for small cylindrical products. The reason for this may be the high powder dispersion and the expansion of certain products when spraying, which causes the layer to loosen due to the shrinkage followed by cooling. APMT sprayed with Nikrothal 80 has shown high porosity in the substrates and high interdiffusion of Fe and Ni. High Ni diffusion in low Al alloys such as K 198 may be a reason why the coated layer cannot optimally form the protective oxide.

• KTH, School of Engineering Sciences (SCI), Mathematics (Dept.).
A Topological analysis of an alternative to the PageRank algorithm in weighted directed graphs2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

In this thesis I use seven data-sets of weighted, directed graphs and present them as weighted, directed k-simplicial complexes. Then, I analyse the topological properties of each data-set in question using Flagser by \cite{flagser} and the cluster at the TU Darmstadt. I then proceed to run two versions of an alternative to the PageRank algorithm. One version contracting and deleting every node visited, and the other deleting only those visited nodes with less than 3 neighbours. I record snapshots of how the data-sets throughout the run of the algorithm, and compute the same topological properties computed before the run. I compare the changes in their homology to understand how the algorithm alters the topology of the graph. I also run several runs of the algorithms to get an idea of how the average graph looks like after the algorithm has been run. I record their new topological properties to find a correlation between the performance of the algorithm and the change in the topology of the graphs.

• KTH, School of Electrical Engineering (EES), Micro and Nanosystems.
KTH, School of Electrical Engineering (EES), Micro and Nanosystems. KTH, School of Electrical Engineering (EES), Micro and Nanosystems.
A Silicon Micromachined 220-330 GHz Turnstile Orthomode Transducer (OMT) in a Low-Loss Micromachining Fabrication Platform2018Conference paper (Refereed)

The work presented in this paper reports on the first wideband OMT in any frequency band implemented by micromachining. This turnstile-junction design provides full waveguide-band operation (220-330 GHz) and is the first implementation of a turnstile-OMT above 110 GHz, since very accurate fabrication is required for this topology. The measured insertion loss is below 0.5 dB and below 0.6 dB for the two polarizations, respectively, with an average measured return loss of 22 dB. Except for some spikes which still are below 30 dB, the cross-polarization is between 50 and 60 dB.

• KTH, School of Architecture and the Built Environment (ABE), Architecture, Urban Design.
Kan utvecklingen av det urbana öka mångfalden av livsmiljöer?2019In: Rurban Planning Talks / [ed] Nils Björling, Mariestad: DaCapo Mariestad , 2019, p. 24-27Chapter in book (Other (popular science, discussion, etc.))

Rurban Planning Talkds är en dialog mellan akademi och planeringspraktik för att klargöra de problem som pågående urbana och regionala omvandlingar medför. Vi vill utveckla och diskutera de samhällsutmaningar, konflikter och möjligheter som uppstår i skärningspunkten mellan landsbygd och stad. Fokus är alla de kommuner som rymmer både landsbygder och städer, och deras gemensamma arbete för hållbar utveckling.

• KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Optimization and Systems Theory.
Clustering Based Outlier Detection for Improved Situation Awareness within Air Traffic Control2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

The aim of this thesis is to examine clustering based outlier detection algorithms on their ability to detect abnormal events in flight traffic. A nominal model is trained on a data-set containing only flights which are labeled as normal. A detection scoring function based on the nominal model is used to decide if a new and in forehand unseen data-point behaves like the nominal model or not. Due to the unknown structure of the data-set three different clustering algorithms are examined for training the nominal model, K-means, Gaussian Mixture Model and Spectral Clustering. Depending on the nominal model different methods to obtain a detection scoring is used, such as metric distance, probability and OneClass Support Vector Machine.

This thesis concludes that a clustering based outlier detection algorithm is feasible for detecting abnormal events in flight traffic. The best performance was obtained by using Spectral Clustering combined with a Oneclass Support Vector Machine. The accuracy on the test data-set was 95.8%. The algorithm managed to correctly classify 89.4% of the datapoints labeled as abnormal and correctly classified 96.2% of the datapoints labeled as normal.

• KTH, School of Electrical Engineering and Computer Science (EECS).
OSM-Based Automatic Road Network Geometries Generation on Unity2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Nowadays, while 3D city reconstruction has been widely used in important topics like urban design and traffic simulation, frameworks to efficiently model large-scale road network based on data from the real world are of high interests. However, the diversity of the form of road networks is still a challenge for automatic reconstruction, and the information extracted from input data can highly determine the final effect to display.

In this project, OpenStreetMap data is chosen as the only input of a three-stage method to efficiently generate a geometric model of the associated road network in varied forms. The method is applied to datasets from cities in the real world of different scales, rendered and presented the generated models on Unity3D platform, and compared them with the original road networks in both the quality and topology aspects. The results suggest that our method can reconstruct the features of original road networks in common cases such as three-way, four-way intersections, and roundabouts while consuming much shorter time than manual modeling in a large-scale urban scene. This framework contributes to an auxiliary tool for quick city traffic system reconstruction of multiple purposes, while there still being space of improvement for the modeling universality and quality of the method.

• KTH, School of Engineering Sciences (SCI), Mechanics.
Enabling high-fidelity measurements of turbulent boundary layer flow over wing sections in the MTL wind tunnel.2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

A reinforced fiber-glass model of a NACA 4412 wing profile is designed and set-up in the Minimum Turbulence Level (MTL) wind tunnel facility at KTH. The model has 65 pressure taps orifices, and the set-up includes two mounting panels designed to allow for particle image velocimetry (PIV) and hot wire anemometry (HWA) measurements of the boundary layer (to be performed in a future campaign). In a first experimental campaign pressure scans are conducted at three angles of attack of interest (5,10 and 12 degrees), and at four different Reynolds numbers based on chord length and inflow velocity (200,000, 400,000, 1,000,000, and 1,640,000). The preliminary results show good agreement with DNS and LES data, however, the effective angle of attack of the wing is affected by the interference of the test section. In order to obtain proper flow conditions for future campaigns inside the test section, wall inserts are designed using 2D k-omega SST simulations. The side-walls are streamlined and the final geometry is corrected to account for the boundary-layer growth over them. The inserts are shown to avoid early separation near the trailing edge at higher angles of attack (10 and 12 degrees), but the 2D simulations fail to capture the aforementioned angle-of-attack issue affecting the pressure distributions. Future extensions of the present insert design should include both 3D simulations of the test-section and a robust optimization procedure to prescribe the resulting pressure distribution.

• KTH, School of Electrical Engineering and Computer Science (EECS).
Machine Learning for Inferring Depth from Side-scan Sonar Images2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Underwater navigation using Autonomous Underwater Vehicles (AUVs), which is significant for marine science research, highly depends on the acoustic method, sonar. Typically, AUVsare equipped with side-scan sonars and multibeam sonars at the same time since they both have their advantages and limitations. Side-scan sonars have a much wider range than multibeamsonars and at the same time are much cheaper, yet they could not provide accurate depth measurements. This thesis is aiming at investigating if a machine-interpreted method could beused to translate side-scan sonar data to multibeam data with high accuracy so that underwater navigation could be done by AUVs equipped only with side-scan sonars.

The approaches considered in this thesis are based on Machine Learning methods, including generative models and discriminative models. The objective of this thesis is to investigate the feasibility of machine learning based models to infer the depth based on side-scan sonar images. Different models, including regression and Generative Adversarial Networks, are tested and compared. Different CNN based architectures such as U-Net and ResNet are tested andcompared as well. As an experiment trial, this project has already shown the ability and great potential of machine learning based methods extracting latent representations from side-scansonars and inferring the depth with reasonable accuracy. Further improvement could be madeto improve the performance and stability to be potentially verified on the AUV platforms inreal-time.

• KTH, School of Electrical Engineering and Computer Science (EECS).
Exploring A Visualization System For History Paths2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Many business intelligence tools aim to digest data into easy, understandable and visualizable information for helping decision-making, while they are still lack of ability to support visualizing the history of selections. This limitation concerns the coming future when everything is about data. Due to it, users are not able to share their thinking paths to the decision. Here a history selection path means a sequence of previous selections. As an approach, it helps users in decision-making and discovery insight. This study investigated an efficient graphical visualization system of history selection paths to support communicating and iterative analysis. We selected tree representation as the main visualization model and also propose features needed for the system.

Specifically, we researched the significance of this study, existing solutions and also the proper designs and functions for the idea. It is initiated by user research including targeting users and scenario mapping. Based on the understanding, we applied a parallel design to narroww down the suitable design. As a result, tree representation was selected as the visualization model. To evaluate whether it touched user needs or not, we applied usability test to collect quantitative data and qualitative comments. For making the test environment as real as possible, a webbased interactive prototype supported by D3.js library was implemented for testing. We analyzed the user experience and also consolidated improvements. As a case study, we implemented the solution on Qlik Sense to verify the possibility to place this solution into real data visualization tool.

Generally, the result of this study formed a valuable initiative for further development and we saw potentials of this tree model system to be used in other areas when it comes to reviewing history as well.

• KTH, School of Electrical Engineering and Computer Science (EECS).
Investigating differences in response time and error rate between a monolithic and a microservice based architecture2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

With great advancements in cloud computing, the microservice architecture has become a promising architectural style for enterprise software. It has been proposed to cope with problems of the traditional monolithic architecture which includes slow release cycles, limited scalability and low developer productivity. Therefore, this thesis aims to investigate the affordances and challenges of adopting microservices as well as the difference in performance compared to the monolithic approach at one of Sweden’s largest banks, SEB - the Scandinavian Individual Bank.

The investigation consisted of a literature study of research papers and official documentation of microservices. Moreover, two applications were developed and deployed using two different system architectures - a monolithic architecture and a microservice architecture. Performance tests were executed on both systems to gather quantitative data for analysis. The two metrics investigated in this study were response time and error rate.

The results indicate the microservice architecture has a significantly higher error rate but a slower response time than the monolithic approach, further strengthening the results of Ueda et. al. [47] and Villamizar et. al. [48]. The findings have then been discussed with regards to the challenges and complexity involved in implementing distributed systems. From this study, it becomes clear the complexity shifts from inside the application out towards infrastructure with a microservice architecture. Therefore, microservices should not be seen as a silver bullet. Rather, the type of architecture is highly dependent on the scope of the project and the size of the organization.

• KTH, School of Electrical Engineering and Computer Science (EECS), Computer Science, Communication Systems, CoS, Network Systems Laboratory (NS Lab).
Universitat Politècnica de Catalunya. Queen Mary, University of London. Independent Researcher. AGH University of Science and Technology in Kraków. Barefoot Networks. University of Vienna.
PURR: A Primitive for Reconfigurable Fast Reroute: (hope for the best and program for the worst)2019In: In International Conference on emerging Networking EXperiments and Technologies, , 2019 / [ed] ACM, 2019Conference paper (Refereed)

Highly dependable communication networks usually rely on some kind of Fast Re-Route (FRR) mechanism which allows to quickly re-route traffic upon failures, entirely in the data plane. This paper studies the design of FRR mechanisms for emerging reconfigurable switches.

Our main contribution is an FRR primitive for programmable data planes, PURR, which provides low failover latency and high switch throughput, by avoiding packet recirculation. PURR tolerates multiple concurrent failures and comes with minimal memory requirements, ensuring compact forwarding tables, by unveiling an intriguing connection to classic string theory'' (\textit{i.e.}, stringology), and in particular, the shortest common supersequence problem. PURR is well-suited for high-speed match\slash action forwarding architectures (e.g., PISA) and supports the implementation of arbitrary network-wide FRR mechanisms. Our simulations and prototype implementation (on an FPGA and Tofino) show that PURR~improves TCAM memory occupancy by a factor of 1.51.5x---10.810.8x compared to a na\"ive encoding when implementing state-of-the-art FRR mechanisms. PURR also improves the latency and throughput of datacenter traffic up to a factor of \mbox{2.82.8x---5.55.5x} and 1.21.2x---22x, respectively, compared to approaches based on recirculating packets.

• KTH, School of Electrical Engineering and Computer Science (EECS), Electrical Engineering, Electric Power and Energy Systems.
KTH, School of Electrical Engineering and Computer Science (EECS), Electrical Engineering, Electric Power and Energy Systems. KTH, School of Electrical Engineering and Computer Science (EECS), Electrical Engineering, Electric Power and Energy Systems. 2ABB Corporate Research, Västerås, Sweden.
Many-objective Optimization of IPM and Induction Motors for Automotive Application2019Conference paper (Refereed)

This paper presents a Pareto-optimality-based optimization methodology suitable for the design of electrical motors in automotive applications. The proposed many-objective evolutionary algorithm is utilized in this study case for the optimization of an interior permanent-magnet (IPM) synchronous motor and an induction motor (IM), considering as criteria the motors' torque capability, efficiency as well as torque density. Finite-element (FE) models of the investigated motor topologies are developed and incorporated in the optimization process in order to ensure an accurate estimation of their electromagnetic performance. The attainment of the targeted specifications by the final optimal designs validates the efficacy of the implemented optimization algorithm.

• KTH, School of Engineering Sciences (SCI), Solid Mechanics (Dept.).
Biomechanics of AAA surveillance patients2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Abdominal aortic aneurysms (AAA) occurs due to local enlargement of the abdominal aorta and affects about 1-2 % of the elderly Swedish population. At admission, AAA patients receive a Computed Tomography-Angiography (CT-A) scanning, and later they are followed-up by 2D ultrasound examinations. There is a need to know if an adequate geometry can be constructed from the baseline CT-A scan and follow-up 2D ultrasounds. To test our hypothesis only CT-A images have been used from six patients and hypothetical ultrasound cross-sections (HUCS) were extracted from the follow-up CT-A scans. The baseline AAA surface was expanded in a mechanical model by applying pressure inside of it until it reached the HUCS. The obtained morphed geometries were then compared with the CT-A-based geometries. The discrepancy between them were calculated by distance measurements. Also, the rupture risk indicators volumes and stresses were compared. Finally, a sensitivity analysis studied the effect of HUCS positioning on the volumes and stresses. The results show that an adequate geometry can be constructed by the investigated concept of morphing. The average distances between the morphed and CT-A-based geometries are 2-4 mm. The average volume difference for the six patients are between 3.8-16.2 %. The wall stress for the morphed and CT-A-based geometries are close only for the first follow-up.

• KTH, School of Architecture and the Built Environment (ABE), Civil and Architectural Engineering, Structural Engineering and Bridges.
Implementation and validation of an isogeometric hierarchic shell formulation2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Within this thesis, thin walled shell structures are discussed with modern element formulationsin the context of the Isogeometric Analysis (IGA). IGA was designed to achieve a directinterface from CAD to analysis. According to the concept of IGA, Non-Uniform RationalB-Splines (NURBS) are used as shape functions in the design and the analysis. Dependingon the polynomial order, NURBS can come along with a high order continuity. Therefore,the curvature of a shell surface can be described directly by the shape function derivativeswhich is not possible within the classical Finite Element Analysis (FEA) using linear meshes.This description of the curvature gives rise to the application of the Kirchho-Love shellformulation, which describes the curvature stiness with the dierentiation of the spatialdegrees of freedom. Based upon this, the formulation can be enhanced with further kinematicalexpressions as the shear dierence vector, which leads to a 5-parameter Reissner-Mindlinformulation. This kinematic formulation is intrinsically free from transverse shear lockingdue to the split into Kirchho-Love and additional shear contributions. The formulation canbe further extended to a 7-parameter three-dimensional shell element, which considers volumetriceects in the thickness direction. Two additional parameters are engaged to describethe related thickness changes under load and to enable the use of three-dimensional materiallaws. In general, three-dimensional shell elements suer from curvature thickness and Poisson'sthickness locking. However, these locking phenomena are intrinsically avoided by thehierarchic application of the shear dierence vector and the 7th parameter respectively. The3-parameter Kirchho-Love, the 5-parameter Reissner-Mindlin and the 7-parameter 3D shellelement build a hierarchic family of model-adaptive shells.This hierarchic family of shell elements is presented and discussed in the scope of this thesis.The concept and the properties of the single elements are elaborated and the dierences arediscussed. Geometrically linear and non-linear benchmark examples are simulated. Convergencestudies are performed and the results are validated against analytical solutionsand solutions from literature, taking into account deections and internal forces. Furthermore,the dierent locking phenomena which occur in analyses with shell formulations areexamined. Several test cases are designed to ensure a validated implementation of the hierarchicshell elements. The element formulations and further pre- and postprocessing featuresare implemented and validated within the open-source software environment Kratos Multi-physics.

• KTH, School of Architecture and the Built Environment (ABE), Civil and Architectural Engineering, Structural Engineering and Bridges.
Design of a low carbon building: Case study of an architectural competition2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Greenhouse gases (GHGs) emissions due to human activities have considerably increased in the past decades which are the main contributors of global warming. In order to limit the consequences of the global climate change happening, all sectors must reduce their carbon emissions and especially the building industry which represents 19% of the carbon footprint of human activities. This paper is giving methods to help reduce the carbon footprint of a building when designing it such as life cycle assessment which allow project teams to compare the global warming potential of all building materials. Those methods are used and challenged in a case study of an architectural competition project named quai d’Issy in Paris, France. Using biobased materials help reduce the carbon footprint of a building, a structure made of timber and concrete elements can emit less than 21% of GHGs than a classic concrete structure. By sourcing reused and recycled building materials, by using geothermal heat pump as heating and cooling systems for example, we have been able to reach for the quai d’Issy project a carbon footprint of 930kgC02eq/m2 of floor area, which is less than level needed for the highest French environmental certification. However, these results can be obtained only if the building materials companies continue their work to develop low-carbon materials and promote recycled and reused materials. This study emphasises the need to spread knowledge of the tools to design low-carbon building to all the actors of the building industry in order to promote behaviours that will limit the consequences of climate change.

• KTH, School of Architecture and the Built Environment (ABE), Civil and Architectural Engineering, Transport planning.
Analysis of Walking and Route-Choice Behavior of Pedestrians inside Public Transfer Stations: A Study on how pedestrians behave in the approaching vicinity of level-change facilities,and how it affects their walking and route-choice behavior2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Pedestrian walking and choice behavior presented was first studied by Fruin in 1971, and since then a lot of research have been carried out in order to understand how humans move and what does make them make choices and obtain certain patterns. In relation to pedestrians, a significant bottleneck inside public stations evaluated by research are the level-change facilities, as Stair Walks and Escalators. The aim of this research is studying how pedestrian behave in the vicinity to stairways and escalators, and how does that affect pedestrian choice, speed and acceleration when choosing one of the two facilities. Also, with a need for more data on pedestrian traffic, further data collection is a big requirement to analyze their behavior and use as tools in future measures. At last, how to optimize the movement of pedestrians in relation to level changes, considering the effects of the movements observed. Two case studies were analysed, Stockholm Central Station and Uppsala Central Station.The study compares data collection methods, tracking methods and previous studies to better fit the scope of this research. The data is backed up from previous research and explains which method better fitted the options available. As a result, video data collection was chosen to collect the data, a semi-automatic tracking software called T-analyst was used to extract speed, trajectories and acceleration from the videos, and microsimulation modelling from VISSIM further investigated different design options to optimize the overall performance and improve travel time in the same area. The analysis found out that there was a possibility to increase the overall performance of the location in higher flow levels, where the most significant queues could be seen, since there was the possibility to achieve higher speeds by modifying the width and position of the stair walks, which allow for a smaller queue in both directions.

• KTH, School of Architecture and the Built Environment (ABE), Civil and Architectural Engineering, Concrete Structures.
Evaluation of air entraining behaviour in concrete using computer aided methods on hardened samples2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Increasing awareness of sustainability in the concrete industry forces structural design and executionto focus on avoiding costly and unpredictable maintenance action, instead paying higher attention todamage prevention by direct actions on early stages of production. One of such approaches, whichdeals with the problem of freezing and thawing deterioration, is intentional air bubble introductionto the concrete mix. However, the mechanism of air entrainment in concrete can be negativelyaffected on different stages of production by many factors including cement type, admixture dosage,casting conditions or mixing procedure. Therefore, reliable tools for the end-product evaluationought to be considered. The experimental study, presented in this work, focuses on understandingthe blast furnace slag (BFS) influence as well as admixtures’ dosage effect on pore structure ofhardened concrete. Three types of cement were evaluated, including ordinary portland cement(OPC) and two types of CEM III cements with different BFS percentage. The optimal amountsof air entraining agent (AEA) and super plasticizer (SP) were chosen and later reduced in orderto evaluate their impact on total air content, spacing factor and specific surface of the air voids.The main method chosen for this evaluation was the use of an office flat-bed scanner to acquireimages and application of BubbleCounter software for the analysis of the air void structure. Thisapproach is based on linear traverse method and requires special surface treatment for contrastenhancement. Specimens for the analysis were cut from hardened concrete cubes and polishedto achieve a flat surface. The samples were later treated with black ink and zinc oxide paste toachieve a clear contrast between white voids and black paste/aggregate area. In order to estimatethe accuracy of this method, more conventional tools such as pressurised gauge method and air voidanalyser were applied for comparison. Resulted mixtures showed significant differences in air voidproperties between OPC and BFS containing concrete, with the latter being less affected by AEAdosage reduction. Changes in spacing factor and specific surface were also registered; however,their deterioration did not follow the same pattern as that of total air content. No significantdeviation between the two cements containing BFS was observed. An interesting effect of the usedpolycarboxylate ether SP on the AEA reactivity was registered, showing deterioration of air voidproperties with the decrease of plasticizer amount. Comparison of the results from different air voidanalysis methods, suggested an overall agreement on the measured air void system changes due tochanging the AEA content. However, the BubbleCounter software tended to slightly overestimatethe material’s resistance to freeze and thaw phenomenon, giving the most optimistic values inspacing factor and specific surface of air voids.

• KTH, School of Industrial Engineering and Management (ITM).
KTH, School of Industrial Engineering and Management (ITM).
Robotization as a driver of increased labour productivity and economic convergence or divergence in the European Union2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

During the years 2004-2014, the manufacturing sector within the EU countries witnessed an increase in the utilization of industrial robots, where robot density per worker approximately doubled. Considering that this is a rather recent event, studies investigating how much industrial robots impact labour productivity are still rare. At the same time, one of EU’s outspoken goals is that of working to foster productivity and economic convergence between the member states. Given the above premises, we have investigated the relation between the adoption rate of industrial robotics within the EU and its effect on labour productivity.

Secondly, we have made a predictive convergence model, in terms of labour productivity. We have collected data from several sources, including the Industrial Federation of Robotics and EU KLMS, in order to build a dataset for our quantitative analysis. We have then used statistical methods such as multiple regressions and 3 stage least square analysis (3sls) to estimate our system of interdependent equations model. The results show that implementation of industrial robotics in the manufacturing sector is a driver of labour productivity. The model finally predicts upward labour productivity divergence between the member states in the years 2015-2025, assuming that the determining factors of labour productivity grow at the same pace in our forecast period as in our data sample.

• KTH, School of Industrial Engineering and Management (ITM).
Utveckling av kursmoment för kemiundervisning: En undersökning av gymnasieelevers kunskapsutveckling i samband med lärandeaktiviteter rörande kemisk analys och hållbar utveckling2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

In this research, the knowledge development of a group of high school students during a course component was studied. The course component, which consisted of four separate parts, was developed for this investigation. The students who participated in the investigation are all in year one in high school, and study at the Science programme. The course component was included in their chemistry course. The course component was developed to increase the students’ knowledge within selected topics from the curriculum, however, also to offer the postulation to develop Scientific Literacy, as well as to increase their ability to discuss societal issues while including a scientific perspective. To be able to carry out the investigation, two different data selections were accomplished. One where the students responded to a survey, where they had to reflect upon their own knowledge development, and another, where their subject knowledge was examined during the assessment of an assignment that they had to carry out as the last part of the course component. The instructions for the assignment was to first describe different analytical tools used in chemistry, and secondly, to narrate different solutions on how to deal with environmental issues, using a perspective of sustainable development. The conclusion drawn from this research is that the course component provided the students the opportunity to increase their knowledge and their strengthen abilities within the requested subject areas. However, it also emerged from the study that, if the different parts om the course component were to be adjusted, the result might have been more successful and the correlation between the course component and the main purposes for it to be carried out might had been stronger. The didactical process that has been carried out during this work has also been analysed. Hence, suggestions about how to reflect upon time requirements and work organisation if one, as a teacher, wishes to either implement this course component in their own tutoring, or do something similar, is presented in the end of this report.

• KTH, School of Industrial Engineering and Management (ITM).
Organizational learning through knowledge sharing:: A study regarding influential factors of knowledge sharing between developers of an IT-consultant organization.2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

This study aims to understand the knowledge sharing process between developers within an IT-consultant organization. As with today’s fast-paced and knowledge-driven society, there’s been a growing interest within organizations to manage their collective knowledge. Thus, this study doesn’t merely focus on what’s currently going on in a specific setting, but also what can be done to further influence the process of knowledge sharing.

Previous research within the field of knowledge sharing has been studied, mainly based upon the theoretical framework developed by Minu Ipe (2003). This study has been further divided into several influential factors regarding the process of knowledge sharing. We, therefore, strive to understand where the act of sharing knowledge occurs, how the employees learn from one another and what motivates them to participate in the process. This was done through qualitative methods, where several employees and a few representatives of the management of a certain organization were interviewed. The results gathered were further analyzed and conclusions were made regarding future improvements to the knowledge sharing process. Social and communicational aspects were determined to be core influential factors of knowledge sharing.

• KTH, School of Industrial Engineering and Management (ITM).
“Factors Influencing FDI Inflows in SouthAsian Countries: A Panel Data Analysis”2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Foreign direct investment (FDI) is played a vital role for boosting up the economies of developing countries. Hence, it is necessary to know the factors that determines the flows of FDI in the developing countries. This study has attempted to investigate how different factors affect the inflow of foreign direct investment in South Asian Countries. To attain the objective this study has collected data on the respective variables for 45 years and considered seven countries. The relationship between different economic variables and their overall impact on FDI inflows have been examined through various panel models like basic pooled OLS estimation, entity fixed effect model, time fixed effect estimation and random effect model. The outcome of this study is that GDP of the country is the main factor behind the FDI inflows in South Asian countries.

• KTH, School of Engineering Sciences (SCI).
KTH, School of Engineering Sciences (SCI).
Interscale transport of Reynolds stresses in wall-bounded flows2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Couette, pipe, channel, and zero-pressure gradient (ZPG) turbulent boundary layer (TBL) flows have classically been considered as canonical wall-bounded turbulent flows since their near-wall behavior is generally considered to be universal, i.e. invariant of the flow case and the Reynolds number. Nevertheless, the idea that large-scale motions, being dominant in regions further away from the wall, might interact with and influence small-scale fluctuations close to the wall has not been disregarded. This view was mainly motivated due to the observed failure of collapse of the Reynolds normal stresses in viscous scaling. While this top-down influence has been studied extensively over the last decade, the idea of a bottom-up influence (backward energy transfer) is less examined. One exception was the recent experimental work on a Couette flow by Kawata, T. & Alfredsson, P. H. (Phys. Rev. Lett. 120, 244501, 2018). In the present work, a spectral representation of the Reynolds Stress transport equation is used to perform a scale-by-scale analysis of the terms in the equation. Two flow cases were studied: first, a Direct Numerical Simulation (DNS) of a Couette flow at a similar Reynolds number as Kawata and Alfredsson. The Reynolds number was ReT = 120, viscosity v. Second, a Large Eddy Simulation (LES) of a ZPG TBL at ReT = 730, 1270, and 2400. For both cases the classic interscale transport or turbulent kinetic energy was observed. However, also an inverse interscale transport of Reynolds shear stress was observed for both cases.

• KTH, School of Industrial Engineering and Management (ITM).
A Novel, Wave-based Control Architecture for Collaborative Haptic Virtual Environments2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Haptic feedback, the introduction of sense-of-touch into virtual environments, has been shown to have many benefits in practical applications. One such area of interest, is the increased performance of tasks that require collaboration. In the literature however, a tradeoff has been identified between transparency and consistency among remotely connected users.

The purpose of this thesis is to explore the realisation of a hybrid network architecture to try to gain the transparency benefits of a peer-to-peer (P2P) architecture, and the consistency and scalability benefits of a client-server (CS) architecture. Unlike the conventional CS case where the client only has a static local copy of the object, the hybrid architecture introduces a dynamics engine on the client side. There is still a central node in the topology, designated the Observer, which contains the central model and all information about the virtual environment. To maintain consistency, the Observer is connected to each client model by a consensus controller, which act as a virtual coupling.

Different distributed control strategies and choice of parameter distributions are investigated and evaluated in simulation. To ensure passivity, and thus stability, wave variable transforms are proposed as an alternative to power variables, which also removes the controller design from the analysis. A passivity-preserving, prediction-based reflection compensation algorithm is also proposed to improve the user experience during collaboration. The proposed solution is generalised for an arbitrary amount of users, for any degree of freedom, while the evaluation is limited to two users in a 1-DOF use-case.

Finally, a comparison is performed in simulation between the proposed hybrid architecture and the state-of-the-art P2P and CS architectures. The findings are evaluated in regards to stability, consistency, and transparency.

• KTH, School of Industrial Engineering and Management (ITM), Machine Design (Dept.), Machine Design (Div.).
KTH, School of Industrial Engineering and Management (ITM), Industrial Economics and Management (Dept.), Industrial Management.
Work functions shaping the ability to innovate: insights from the case of the safety engineer2019In: Cognition, Technology & Work, ISSN 1435-5558, E-ISSN 1435-5566Article in journal (Refereed)

To perform as intended, firms are divided into work functions that contribute to the behaviour-shaping constraints under which individuals build their skills, knowledge and networks. These in turn provide a specialized perspective on organisational structure and culture. In a mixed methods study involving interviews and statistical analyses, we investigate whether a work function can thereby affect individuals’ innovation foci and efficacy. Safety engineers, central to firms developing Cyber-Physical Systems (CPS), are shown to play a significant role in identifying and supporting viable innovation related to organisational aspects. Results indicate that safety engineers could use the firm’s collaborative innovation process to facilitate organisational learning, helping different work functions come together to construct working practices aligned with the organisational values of the firm. This is explained by their understanding of safety culture—a specialized understanding of organisational culture and complexity provided by their work function. We conclude that work functions that provide their members with a perspective well aligned with their firm’s organisational values can instil the ability to identify and support organisational innovation. This suggests that safety engineers in CPS domains could be effective in a mediatory role, facilitating innovative changes to organisational structures and processes when introducing and operating safety management systems. Stronger aspects of organisational liaising and systems thinking could reinforce this ability—the former by a wider scope and motivation leading to an increased skill in communicating with dissimilar individuals, and the latter by providing the skills and tools needed to analyse the politics of complex organisations.

• KTH, School of Electrical Engineering and Computer Science (EECS), Intelligent systems, Micro and Nanosystems.
KTH, School of Electrical Engineering and Computer Science (EECS), Intelligent systems, Micro and Nanosystems. KTH, School of Electrical Engineering and Computer Science (EECS), Intelligent systems, Micro and Nanosystems.
Dielectric Rod Antenna Array for Photonic-Based Sub-Terahertz Beamforming2019Conference paper (Refereed)

This work presents a dielectric rod antenna array designed for a photonic-enabled beamforming system at subterahertz frequencies. The photonic chip generates an optical group delay, providing the beam-steering capability. The antenna array is fabricated from high-resistivity silicon by micromachining. Simulation results demonstrate a directivity of 14.7 dBi and a beam steering range of 56 degrees at 85 GHz. The system is intended as a sub-THz transmitter for broadband wireless communications.

• KTH, School of Electrical Engineering and Computer Science (EECS).
Generative Adversarial Networks in Text Generation2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

The Generative Adversarial Network (GAN) was firstly proposed in 2014, and it has been highly studied and developed in recent years. It has obtained great success in the problems that cannot be explicitly defined by a math equation such as generating real images. However, since the GAN was initially designed to solve the problem in a continuous domain (image generation, for example), the performance of GAN in text generation is developing because the sentences are naturally discrete (no interpolation exists between “hello" and “bye").

In the thesis, it firstly introduces fundamental concepts in natural language processing, generative models, and reinforcement learning. For each part, some state-of-art methods and commonly used metrics are introduced. The thesis also proposes two models for the random sentence generation and the summary generation based on context, respectively. Both models involve the technique of the GAN and are trained on the large-scale dataset. Due to the limitation of resources, the model is designed and trained as a prototype. Therefore, it cannot achieve the state-of-art performance. However, the results still show the promising performance of the application of GAN in text generation. It also proposes a novel model-based metric to evaluate the quality of summary referring both the source text and the summary.

The source code of the thesis will be available soon in the GitHub repository: https://github.com/WangZesen/Text-Generation-GAN.

• KTH, School of Electrical Engineering and Computer Science (EECS).
FoodEd: Encouraging sustainable food waste practices through an informative app2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Every year 1.6 billion tons of food is thrown away corresponding to one third of all the food that is roduced. It accounts for 8 % of the global greenhouse gas emission, meaning that it has a huge negative impact on the environment. In Sweden, households account for the largest amount of food waste. Therefore, this study focuses on reducing food waste on a household level in Swedish households.

There is a growing interest in HCI for how design can be used for reducing food waste by changing unsustainable behaviour. In this work a working mobile prototype, FoodEd, was designed, implemented and studied. FoodEd contains information of how to handle and store different food items in an optimal way. The aim was to raise awareness of good food handling practices and edibility beyond expiry dates in order to prevent unnecessary food waste in households.

In this study, FoodEd was used by 10 participants during a week’s time. Two qualitative interviews were held with each participant during the study. The findings of the study indicate that an informative design can be effective in providing relevant information for its users leading to more sustainable food waste behaviours.

• KTH, School of Electrical Engineering and Computer Science (EECS).
An Explorative Study of Interaction with Tracked Objects in a Virtual Reality Game2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

In recent years, Internet enabled objects (Internet of Things) and their augmentation through virtual reality has become both technically possible and increasingly advanced. This paper explores how interaction between physical IoTs and their virtual and digital twins can be advanced. Three virtual reality games focusing on three different interactions were created. One game focused on rotating, another on pushing and pulling, and a third on lifting and dropping. All games revolved around the use of real tracked cardboard boxes which were represented in the games by a virtual box in the same relative position and rotation to the player, allowing participants to manipulate the virtual boxes by manipulating the physical boxes. 14 participants were asked to play the games and were afterwards interviewed in regards to the games, their interactions with them, their enjoyability and their thoughts regarding them and the concept in general. The results were acquired from game performance, the reflections of the participants, and the subsequent analysis of recorded audio and video. The study presents the difficulties, challenges and opportunities of such a system, while also providing insight into lessons learned from the creation of the system and the games. The main contributions of the paper are the lessons learned in creating the games and experiences in addition to a few specific areas of interest for future research on the area, namely the importance of ergonomic consideration and affordance evaluation.

• KTH, School of Electrical Engineering and Computer Science (EECS).
Key Tension Points of creative Machine Learning applications keeping a Human-in-the-Loop2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

Machine learning (ML) and artificial intelligence (AI), might have earlier primarily found industrial use, improving production chains, efficiency and the like but are now an integral part of private and commercial application. Many systems are using, or are claiming to use, machine learning to improve the end user's experience.

This study aims to explore applications that are using creative ML, in which output might have a plethora of solutions instead of a single correct one. More specifically the focus is to evaluate which Key Tension Points, central lesser components of a complex and bigger issue, arise for researchers, designers and users coming into contact with this technology.

The goal is to draw upon these Key Tension Points and attempt to frame guidelines which researchers and designers can use to further their understanding of the relationship between ML and design and how they can be accounted for to build and develop better application. The study found that these tension points (Impersonality, passive consumers & transparency) vary considerably depending on the application and presents how designers can account for them.

• KTH, School of Electrical Engineering and Computer Science (EECS).
Framgångsfaktorer inom digital livsmedelshandel: Samsyn och skillnader mellan konsumenter och aktörer2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

In this research the differences and consensus between the consumers and the actors perception of the crucial success factors within Swedish online grocery shopping are being analyzed.

Online grocery shopping is one of the fastest growing markets in Sweden. The online grocery shopping market encompasses both pure digital players and major hybrid actors with a combination of online and offline sales.

The main result in this report show that consumers and actors agree on some of the main issues even though the actors seems to overrate the importance of making the webpages personalized with each customers favorites. At the same time they seems to underrate the importance of keeping the price at the same level as physical stores and the importance of favorable membership. There is consensus on the importance of an easy-to-use webpage and home delivery being the crucial form of delivery.

It is of greatest importance that the actors develop strategies that make them relevant and attractive to consumers, competitive in the market and also profitable. The market is in a very interesting phase right now, experiencing high growth and there is a lot of potential for successful actors. Big strategic decisions and large investments are being made, but as this research suggests not always on things that really could make a difference for the consumers and could make them loyal to one actor and give that actor a greater part of their spend.

• KTH, School of Industrial Engineering and Management (ITM).
Initial Assessment ofManufacturing ExecutionSystems: Development of a methodology to definebusiness needs and functional requirements2019Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis

A key component of the smart factories that industry 4.0 introduce is the MES (manufacturing executionsystem). The MES lies above the actual shop floor in the enterprise system hierarchy and is not in directcontact with the actual production, in the way PLC (programmable logic control) or SCADA (supervisorycontrol and data acquisition) systems are. MES guide, initiate, respond to and reports on the productionactivities as well as distribute information to other company IT systems, bridging the gap between thecontrol systems of the shop floor and the information systems on an enterprise level. MES is a class ofinformation systems built to support shop floor processes and improve the integration to other departmentsof the enterprise by incorporating these systems into the overall IT-architecture. The main goal of an MESis to improve and optimize the production management functions and increase visibility into themanufacturing process.

The role of the MES is defined by industry standards that identify what functionalities and dataflows anMES should include, and how it is intended to be integrated with other information systems. However, littleresearch has been focused on adopting these standards in actual MES implementation projects. The MESA(manufacturing executions system association) is presenting a standardized definition of MES as well as 2models of the, according to the standard, 11 key MES functionalities. The ISA-95 (IEC/ISO 62264)standard identifies sub-systems of an MES and defines the boundaries between the ERP, MES and otherautomation and IT systems. Company parameters such as manufacturing environment, production modeland plant type all affect the business needs and what MES functionalities are of priority. Hence, for an MESimplementation project, a business-specific evaluation must be performed.

Prior research in the area is presenting a high-level workflow and best-practices of an MES implementationproject. By combining this workflow with the general software implementation standard ISO/IEC 12588(ISO/IEC 15288: System engineering – System lifecycle process), a methodology for performing the initialassessment of a company’s MES needs and business requirements is built. In the methodology models fromMESA and ISA-95 are applied to ensure an industry-accepted terminology and process. The objective ofthe methodology is to provide a standardized way to make an initial assessment of a company’s MES needsand specify system requirements. The methodology is validated through a study performed at discretemanufacturing line.

The overall needs and specific functional requirements are identified through the methodology and arepresented according to a URS (user requirements specification) for an MES. The requirements areprioritizing according to MoSCoW analysis. Additional validation of the methodology must be performedto further evaluate the suitability of using the methodology for initial assessments of businesses’ MES needs.