Chalmers Open Digital Repository
Välkommen till Chalmers öppna digitala arkiv!
Här hittar du:
- Studentarbeten utgivna på lärosätet, såväl kandidatarbeten som examensarbeten på grund- och masternivå
- Digitala specialsamlingar, som t ex Chalmers modellkammare
- Utvalda projektrapporter
Forskningspublikationer, rapporter och avhandlingar hittar du i research.chalmers.se
Enheter i Chalmers ODR
Välj en enhet för att se alla samlingar.
Senast inlagda
Multivariate anomaly detection with LSTM layered Variational Autoencoder
(2025) Blixt, Emily; Kullmyr, Linus
The aim of this thesis was to develop and evaluate the effectiveness of a recurrent neural
network layered autoencoder model for detecting anomalies in multivariate time-series
data, with a focus on improving the accuracy and reliability of diagnostic data for Volvo
Penta’s boats. The primary goal was to leverage the relationships and correlations between
signals to identify deviations that traditional models may fail to detect. The
model’s performance was assessed in terms of its ability to learn the structure of normal
data, detect synthetic anomalies, and provide meaningful insights without relying on labeled
datasets.
The study highlights the limitations of traditional evaluation metrics, which are often unsuitable
for unsupervised learning approaches like the model used. Instead, the model’s
effectiveness was demonstrated through reconstruction error analysis and its ability to
handle the complexities of multivariate time-series data. Challenges such as data dimensionality,
sequence length optimization, and noise handling were addressed to enhance
the model’s robustness. The findings suggest that while the model excels at identifying
synthetic anomalies and capturing temporal relationships, further work is needed
to generalize its capabilities to real-world scenarios. This research lays the groundwork
for improving diagnostic processes and supports the development of more adaptive and
reliable anomaly detection systems.
Camera simulation tool for automotive applications
(2025) Smajic, Amir; Kamal Mohammad, Bawan
As technology in the automotive industry consistency evolves, and driver aids get more advanced the need for advanced driver assistance systems (ADAS) grows larger. The need for quicker and more reliable simulation tools becomes essential to further evolve ADAS. This thesis presents the design and development of a simulation tool for 3D camera position testing. The tool created using blender and blender’s own Application Programming Interface (API) and blenders own node functions to create a realistic environment to improve performance in the sense that it produces a reduced gap in between simulation and actual performance. The simulation tool was created in collaboration with Polestar to create an improved simulation tool to an already existing version with the limitation of the previous tool being setup-time and a steep learning curve. By creating a personalized interface with real-time visualization before rendering set in a realistic environment and with adjustable camera settings the simulation tool. The tool enables engineers to seamlessly integrate CAD models into the simulation tool without adjusting coordinates. The finished tool was tested by Polestar’s ADAS team and found the tool to be both smoother while using the tool and a more efficient setup time which minimized time consumption
Förankrad frontregel i flervåningsbostadshus med trästomme: En studie av träets formförändringar och dess påverkan på förankringssystem i träkonstruktioner
(2025) Hafstrand, Theo; Öman, Samuel
I detta examensarbete undersöks problematiken med förankringen av Deromes flervåningsbostadshus med trästomme. När vindlaster verkar på konstruktionen uppstår vertikala lyftkrafter i väggarna. I nuläget hanterar Derome dessa lyftkrafter genom att förankra byggnadens fronteglar med hjälp av stålstag mellan bottenplattan och bjälklaget. Problematiken med att kombinera trä och stål över långa avstånd är att skillnaderna i materialegenskaper medför betydande konsekvenser för förankringen, främst till följd av träets fuktrelaterade rörelser och krypningsbeteende.
Examensarbetet syftar till att analysera ett konceptuellt alternativ till Deromes nuvarande lösning i form av ett fjädrande vajersystem. För att förstå vilka tekniska faktorer som påverkar förutsättningarna för förankringen att verka på önskat sätt samt vilka krav som ställs på en lösning genomförs litteraturstudier, beräkningar och konceptuell analys.
Genom den konceptuella analysen framgår det att både fjädrande och styva förband endast har optimal verkan vid ett visst läge i träets fuktrörelsecykel. Det framtagna konceptet uppfyller därmed inte de satta kraven samtidigt, men ger förståelse för kärnan i problemet och ger insikter för vidare forskning. Vidare visar det sig att infästningar över kortare avstånd inte lider av denna problematik och antyder på att en annan designstrategi kan vara mer lämpad.
Mitigating label noise in ECG data: A comparative analysis
(2025) Manouras, Manousos; Pediaditis, Dimitrios
Label noise in electrocardiogram (ECG) datasets, where samples are incorrectly labelled, significantly hinders the performance of machine learning models by fitting to the incorrect labels. This type of noise can arise from several factors, such as human error, inter-expert variability, or obsolete automated annotation algorithms, leading to inconsistencies within dataset labelling. In this thesis work, three noise
mitigation methods are compared with a baseline model to evaluate both the impact of label noise and the effectiveness of these mitigation strategies in ECG datasets. The mitigation methods chosen are Stochastic co-teaching, Self-learning and DivideMix. Class-dependent label noise was synthetically introduced into two ECG datasets, PTB-XL and CODE15%, comprising of symmetric and asymmetric noise types with rates of 20% and 40%. The best-performing method, as quantified by the AUROC score, was self-learning, with improvements from 4 to 8% over the baseline in CODE15% and from 8 to 12% in PTB-XL. DivideMix demonstrated reduced performance, presumably because it has been optimised for specific image datasets. Stochastic Co-teaching achieved better results on the CODE15% dataset, likely due to the larger sample size of this dataset. Furthermore, an additional ECG dataset obtained from Akershus University Hospital was used to assess the generalisability of the best-performing method under unknown noise conditions. The results did not show an improvement over the baseline model, indicating a strong dependency between the characteristics of the dataset and the effectiveness of noise mitigation strategies.
The role of suppliers in the market of reused construction materials A supply chain perspective
(2025) Nilsson, Emma; Runesson, Olle
The construction industry is the industry that is using the largest amount of natural
resources. The model of the construction industry has for decades been to “take,
make, use, dispose”, a linear model not allowing materials in properties to be reused.
Changing the construction industry to a circular model including reuse of construction
materials is a way to reduce the impact on the climate from the construction industry.
The suppliers of construction materials possess an important role in the
transformation of scaling up the market of reused materials. Thus, the purpose of this
thesis is to identify and analyze the Supply Chain role of suppliers in the market of
reused construction materials. The research is based on interviews with 29 actors in
the construction industry. First, actors involved in three projects were identified and
interviewed, to create a holistic view of how the market of reused materials function.
Thereafter, other actors, for instance suppliers, were interviewed to create an
understanding of how different stakeholders in the construction industry work with
reused materials.
The interviews resulted in information about three projects and the processes of
working with reused materials from a supplier’s and reuse hub’s point of view.
Furthermore, it contributed with information about market drivers for the reuse
market, challenges, collaboration and opinions about the future of the reuse market.
Seven actors were identified as included in the market of reused materials. These are
property owner, construction company, reuse consultant, supplier, reuse hub,
architecture firm and demolition firm. Furthermore, several resources controlled by
the actors and activities performed by them were identified.
A number of key factors influencing the reuse market is discussed in the thesis. These
factors are environmental sustainability, costs, market requirements, collaboration,
communication, procurement process, logistics, matching supply and demand, and
supplier capacity.
The result from the study implies that suppliers possess an important role in the
transformation of the reuse market. There are four main activities suppliers should
implement to establish work with reused materials. First, they need to adapt
operations to enable circular flow. Secondly, they need to focus on the take back
process. Furthermore, the process at the suppliers must be updated, and finally the
selling process of reused materials needs to be implemented.