Chalmers Open Digital Repository
Not a member yet
26247 research outputs found
Sort by
Enhancing Gas Adsorption In Sensors：Au–Pt Nanoparticles and Methylated Coordination Cages
No full textThis thesis investigates two nanostructured materials, bimetallic gold platinum (AuPt) nanoparticles and a methylated cobalt(II)/iron(II) coordination cage, for their potential use in gas sensors designed to detect acetone, a key biomarker for
non-invasive glucose monitoring. The AuPt nanoparticles were synthesized using a modified co-reduction method based on Britto’s procedure. The synthesis was optimized to achieve uniform and well-dispersed nanoparticles through precise control of reaction time and washing steps. The nanoparticles were characterized using SEM and, EDX, to confirm their morphology, composition, and crystal structure. They were immobilized on quartz
substrates through silane functionalization, and UV–Vis spectroscopy was used to verify both immobilization and acetone vapor adsorption. A clear spectral shift observed after acetone exposure confirmed successful adsorption on the nanoparticle surface, indicating their suitability for gas sensing applications.
In parallel, methylated Co(II) and Fe(II) coordination cages were synthesized through subcomponent self-assembly and characterized using NMR and UV–Vis spectroscopy. The immobilization of Co(II) cages on glass and silicon substrates was verified using UV–Vis and TOF-SIMS. The data is consistent with successful immobilization of the cages on both glass and silicon surfaces. Notably, this study demonstrates for the first time that this particular cage has been successfully adsorbed onto glass and silicon substrates. Gas exposure experiments were performed on the Fe(II) cage solutions and analyzed by NMR; however, no significant spectral changes were detected, likely due to gas dissipation or weak interactions at ambient conditions.
Overall, the study presents promising results for the AuPt nanoparticles, while preliminary findings for the coordination cages indicate the need for further optimization. These results contribute to the development of nanostructured mate-
rials for more sensitive and selective gas-sensing technologies aimed at advancing non-invasive diagnostic methods
Leveraging AI to Enhance Innovation Efficiency in the MedTech Sector AI Applications and Strategies for Enhancing Innovation Efficiency: A Case Study in the MedTech Sector
No full textThe MedTech industry is experiencing rapid technological change and facing evolving
healthcare needs, which puts pressure on actors to innovate efficiently and stay competitive.
At the same time, Artificial Intelligence (AI) is emerging as a transformative tool with the
potential to enhance efficiency. However, despite its potential, the adoption of AI within the
MedTech sector remains limited. The aim of the thesis is therefore to examine how MedTech
companies can leverage AI to enhance innovation efficiency. The main research question is
explored through the lens of an innovation process framework, consisting of the three phases:
Idea, R&D, and Commercialization. Within this framework, four key areas have been
identified with the potential to improve overall efficiency by the integration of AI, namely:
Knowledge Management, Patent Analytics, Market and Customer Analysis, and Resource
Allocation. In addition, the study addresses important aspects of implementing AI within
organizations to optimize the possible outcomes.
A single case study has been conducted analyzing a MedTech company, which was selected as
it constitutes a suitable case to explore the research question. The methodology consists of a
literature review followed by an interview study. The literature presents the current state of AI
applications in the four focus areas defined and reviews opportunities and challenges
regarding implementation. The interview study included participants both working at the
company studied, to depict its current practices and needs, and individuals at other
organizations who are at the forefront of AI adoption, to gain best-practice perspectives and
expert insights.
The results show that AI technologies have great potential in enhancing innovation efficiency
in the MedTech industry. AI helps to improve the flow of information inside the organization,
strengthen intellectual property strategies, enhance market and customer insight, and optimize
resource allocation between projects. However, to achieve those benefits, a highly coordinated
strategy with top-down support is necessary. Thus, the thesis concludes that a structured and
human-centered approach to AI adoption is essential for companies in the complex and highly
regulated MedTech landscape to remain competitive and achieve long-term innovation
success
Design och implementation av ett GPU-programspråk med högre ordningens funktioner
No full textA new programming language, gpulang, is introduced together with its abstractions that enable automatic parallelisation of programs. Central to the language is the concept of streams, an abstract data type that may only be inspected through built-in higher-order functions such as map, reduce, and filter. This restricted access model is intended to facilitate optimisation by the compiler. We describe the design of the language as well as the implementation of its compiler, which includes code generation for NVIDIA and AMD GPUs using the existing HIP framework. Preliminary performance measurements suggest that gpulang achieves performance comparable to the functional GPU language Futhark, but trails behind the Thrust C++ library
Extremvärdesanalys av nederbördstrender i Sverige
No full textDenna rapport undersöker trender i extrema dygnsnederbördshändelser vid 37 svenska mät stationer med syfte att avgöra om extrem nederbörd blivit intensivare och frekventare i takt
med klimatförändringarna. Datan som användes är över dygnsnederbörd och hämtades från
SMHI. Den analyserades med extremvärdesteori i programmeringsspråket R, det tillämpades
huvudsakligen två metoder, tröskelmetoden och block maxima-metoden tillsammans med en
GEV-fördelning och en GP-fördelning. För tröskelmetoden analyserades överskridandens av stånd till 99.5% kvantilen och för block maxima-metoden analyserades årliga maximum. Som
komplettering gjordes även en modellering med en inhomogen Poissonprocess för att under söka frekvens. För samtliga analyser användes likelihoodkvot-test varefter p-värden beräknats
för att bedöma signifikans, med signifikansnivå α = 0, 05. Resultatet visade att 16% av sta tionerna hade signifikant trend enligt block maxima-metoden, medan 10% hade signifikans
med tröskelmetoden. Den inhomogena Poissonprocessen ger däremot ett tydligare resultat,
där uppvisar 38% av stationerna signifikanta trender. Baserat på medianen av trendskatt ningarna i frekvens har det skett en ökning med 55% de senaste 60 åren. Detta tyder på att
extrem nederbörd har blivit mer frekvent, även om de årliga maximumen inte ökat i samma
uträckning. Det i sin tur tyder på att den statistiska osäkerheten var för stor för att upptäcka
alla trender i årliga maximum. Avslutningsvis tas samhälleliga aspekter och konsekvenser av
ökad extrem nederbörd upp, såsom påverkan på infrastruktur och översvämningsrisker. Det
ges också förslag på framtida projekt. Bland annat hade det varit intressant att studera hur
nederbörden ökar med avseende på temperatur
A New Dawn for the Heavy Truck Dealer Network Opportunities related to increasing electrification and direct sales
No full textFor intermediaries to stay relevant in distribution networks, they need to adapt to
changing conditions. Currently, heavy-duty truck manufacturers have committed to
increase their offering of electric vehicles to minimise the environmental impact of
transport. At the same time, firms are developing their distribution structures and
strategies with the aim of fulfilling customer requirements and reducing costs, which
implies utilising direct sales. The aim of this master’s thesis is to investigate future
business opportunities for an intermediary in a distribution network of heavy-duty
trucks, in a scenario with increased electrification and direct sales. This master’s thesis
is based on a single case study in collaboration with Volvo Group at Group Trucks
Technology, focusing on the internal and private dealers of Volvo Trucks’ heavy-duty
truck dealer network. To fulfil the aim of this thesis and address the research problem
dealt with in this study, several issues have been investigated. The current operations
of the dealers have been studied, with the aim of identifying core activities and
resources to outline the current dealer roles. Taking the current dealer roles and
operations as a point of departure when analysing the impact of increased electrification
and direct sales, implications for the future dealer role could be outlined. Increased
electrification and direct sales are reshaping the business environment for the dealers,
driving significant changes in current activities, resources, and roles. Increased
electrification and direct sales imply a reduction in workshop service and sales
opportunities, alongside the increased importance of supporting customers in the
transition to electric vehicles. Continuing to be a main customer contact point, the future
dealer role will be more complex, moving towards an advisory role emphasising the
need to provide the customers with comprehensive solutions. The study shows that the
dealer roles are being redefined rather than eliminated. For the dealers, this implies that
they need to consider their roles to ensure that they will continue to be relevant actors
that create value in the distribution network. Future business opportunities for the
dealers are presented in response to the impact of increased electrification and direct
sales
UFRM: A User Feedback Reference Model for Managing Feedback in Dynamic Software Scenarios - Systematic Approach to Developing and Evaluating the UFRM in Dynamic Scenarios
No full textUser feedback is essential for software improvement, shaping usability, functionality, and overall user experience. However, in Dynamic Scenarios, where users try to provide feedback under high cognitive load, stress, or mental pressure due to factors such as time sensitivity, environmental uncertainty, or task-focused workflows, significant challenges arise. These situations make it difficult for both feedback senders and receivers to effectively manage the feedback process. As a result, users may delay or skip giving feedback altogether, making it harder for receivers to collect and process valuable input, especially in dynamic scenarios.
This study develops a User Feedback Reference Model (UFRM) to improve feedback management under such conditions. We conducted 10 semi-structured interviews with feedback receivers (such as developers and product managers) and 30 semi-structured interviews with feedback senders (end users) to understand their current issues in collecting and processing feedback in dynamic scenarios. The results were analyzed using Thematic Analysis to identify key preferences and challenges of both sides and suggest solutions. These findings were then structured into a four-layer conceptual model for managing real-time feedback in dynamic scenarios, covering scenario detection, feedback collection, processing, and response.
UFRM was validated using three real-world inspired use cases covering smart navigation systems, autonomous autopilot driving, and digital healthcare platforms to ensure its effectiveness. These use cases were derived from real feedback contexts shared by users during interviews and were applied step-by-step to test the model’s adaptability and performance under realistic dynamic constraints. The findings provide insights into optimizing feedback mechanisms in dynamic scenarios, balancing the preferences of both feedback senders and receivers, and supporting better software adaptation and user experience
Motverkan av isbildning och avisning av drönar-propellrar
No full textIn sea rescue operations performed by Sjöräddningssällskapet swedish sea rescue, having an almost immediate evaluation of the scope and nature of an accident is critical in making informed decisions. This evalutation is essential for making decisions regarding the allocation of resources, such as equipment, personnel and rescue vehicles. Sjöräddningssällskapet makes use of drones to rapidly gather visual information at the accident in favorable weather conditions. Typically the drone can reach the scene within five to twenty minutes at its maximum speed when the ambient temperature is above zero degrees Celsius. However, in atmospheric icing conditions the moisture in the air creates ice buildup on the propeller causing the drone to operate well below its optimal capacity. Leading to reduced flying performance due to the propellers disrupted aerodynamics and increased weight due to ice, with a higher energy consumption.In the worst case scenario the drone is not able to reach the accident site before the battery is depleted. Therefore it is necessary to develop a system to counteract atmospheric icing conditions during the drone rescue operations, ensuring reliability in performance and functionality of the drone.
The aim of this thesis is to investigate and develop methods to avoid and reduce the ice buildup on the propeller blade. With a focus on minimizing flight time reduction and energy draw. Keeping the reductions to a minimum to maximise the drones distance and speed while avoiding ice accretion. This study contains reviews of both currently used deicing techniques and their suitability to this small scale usage and more experimental techniques not in use for deicing applications. To identify promising and optimal solutions that can be further looked into. The finding can be utilised to improve the performance of drones during sub-zero weather, therefore improving search and rescue operations over sea
Adaptive KV Cache Management for Efficient Transformer-based LLM Inference - Leveraging Attention Sparsity for Memory Optimization
No full textThis Master’s thesis addresses the critical challenge of memory inefficiency in Transformerbased Large Language Models (LLMs) during inference, specifically focusing on the prohibitive memory footprint of the Key-Value (KV) cache. As LLMs scale, the KV cache becomes a significant bottleneck, limiting longer context windows and overall operational efficiency. To mitigate this issue, we propose and evaluate Adap-KV, a novel adaptive memory management strategy for the KV cache. Adap-KV employs a layer-aware dynamic allocation approach that intelligently adjusts KV cache size in real-time, leveraging insights from attention sparsity patterns. Our method aims to optimize memory utilization without compromising the performance or quality of LLM inference. Experimental results demonstrate that Adap-KV significantly reduces KV cache memory consumption, thereby enhancing the efficiency and scalability of Transformer-based LLMs, making them more amenable for real-world deployments with extended context capabilities
Sjukhusvård i hemmet för behandling av akut hjärtsvikt En intervjustudie om barriärer och åtgärdsförslag vid implementering av en ny vårdmodell
No full textBehovet av fler vårdplatser i Sverige ökar som svar på en åldrande befolkning.
Däremot har det på grund av bland annat resursbrist antalet vårdplatser istället
minskat. Hospital at Home (HaH) är ett vårdkoncept som kan avlasta
akutmottagningar då patienten vårdas i sitt hem med hjälp av egenmonitorering.
HaH har potential att implementeras för patienter med hjärtsvikt, en vanlig
folksjukdom bland äldre i Sverige. Däremot finns det barriärer för att implementera
HaH i svensk hälso- och sjukvård. En vårdmodell framtagen av en projektgrupp vid
Sahlgrenska Universitetssjukhuset har varit en central del i denna studie som syftar
till att identifiera barriärer för implementering av denna vårdmodell samt att föreslå
åtgärder för att hantera barriärerna.
Forskningsstudien genomfördes induktivt och kvalitativ data samlades in via en
intervjustudie, auskultering samt en workshop. Utifrån intervjustudien identifierades
35 aspekter med inverkan på implementeringen av vårdmodellen. En efterföljande
tematisk analys av dessa faktorer resulterade i att 13 specifika barriärer för
implementering kunde urskiljas. Under en workshop validerades barriärerna och
prioriterades i ett PICK-diagram. Barriärerna Olika journalsystem (barriär 2), Otydlig
ansvarsfördelning (barriär 11), Informationsbrist (barriär 12) och Direktinläggning ej
möjligt (barriär 13) klassades som de största utmaningarna för implementeringen av
vårdmodellen.
Slutsatsen drogs att merparten av de identifierade barriärerna orsakades av att
organisationens formella strukturer inte är anpassade till den tilltänkta vårdmodellen.
Kortsiktigt kan andra delar av organisationen eller platsanpassade lösningar
kompensera för detta men långsiktigt behöver organisationens formella strukturer
förändras innan implementation sker