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
 

Enheter i Chalmers ODR

Välj en enhet för att se alla samlingar.

Senast inlagda

Post
AI-Based Wireless Channel Prediction - Generalized Models for Adaptive Measurements and Predictions
(2024) Chen, Bingcheng; Chalmers tekniska högskola / Institutionen för elektroteknik; Svensson, Tommy; Winges, Johan; Kumar Nagalapur, Keerthi; Sattari, Mehdi; Zhang, Xinlin
In wireless communication systems, channels can be time-varying, and their conditions can change due to factors such as mobility, interference, and environmentalconditions, which brings challenges in maintaining communication reliability. Hence, acquiring channel state information (CSI) is a critical step in physical layer of wireless communication. However, when accurate CSI or precoder fed back by a User Equipment (UE) is used by a network for MIMO precoding, the received CSI/precoder can get outdated quickly, consequently resulting in a loss of user and system throughput due to the utilization of outdated precoders. In a feature introduced in 3GPP Rel-18, a UE can be configured to measure multiple instances of the channel, use them to predict a number of channel/precoder instances in the future and report the predictions to the network. The performance of such a scheme depends highly on the ability to predict the channel with high accuracy. In this study, we explored AI-based methods for channel prediction to address this challenge. By utilizing a past window size of measured channel sequences, the proposed method forecasts future channel conditions. Out of all the AI models tested, the Transformer Encoder-Only model demonstrated superior performance, its counterparts and a classical non-AI based autoregressive (AR) model. We also found that, training the model with a diverse dataset with a mix of UE velocities, embedding across the time dimension within the Transformer EncoderOnly model, and constructing the model within the complex domain yields enhanced generalization capability. Furthermore, we developed a generalized model capable of using varying number of channel measurements to predict varying number of channel instances in the future.
Post
Holistic Diagnosis via Multimodal Foundation Models
(2024) Pauli, Oskar; Chalmers tekniska högskola / Institutionen för elektroteknik; Graell i Amat, Alexandre; Ceccobello, Chiara; Östman, Johan
The healthcare domain has data in many different forms, or modalities. They can be in the form of x-ray images, time-series of certain events like heart rate or blood pressure, textual data from notes etc. Medical practitioners uses many different modalities every day to make informed and sound decisions. With the recent success of small and large language models, it is natural to try and incorporate them with multimodal capabilities in the healtcare domain. This thesis seeks to investigate how well small language models can perform on predictive tasks in healthcare using multimodal data. To explore this, projectors that project data from different sources to the embedding space of a language model was developed. While the results show that a multimodal language model is better than a single-sourced version, it is still being outperformed by the XGBoost model. Even though it is being outperformed, the model proposed shows promise in regards to generalizability, potentially streamlining predictive tasks in healthcare. The thesis argue that even if improvements needs to be made and the challenges it poses can be difficult to handle, further advancements can lead to facilitating medical practitioners in a very efficient way.
Post
Tidig karaktärisering av stroke genom videoanalys, maskininlärning och ögonspårning
(2024) Ollila, Samuel; Ström, Eddie; Khatiri, Robin; Svensson, Teodor; Westerberg, Jacob; Chalmers tekniska högskola / Institutionen för elektroteknik; Chalmers University of Technology / Department of Electrical Engineering; Canderfjord, Stefan; Jalo, Hoor
Stroke är en ledande orsak till dödlighet och funktionsnedsättning globalt. Snabb och tillförlitlig diagnos är avgörande för att optimera behandlingen, öka patientsäkerheten och rädda liv. Detta projekt syftar till att använda maskininlärningsalgoritmer för att bedöma misstänkta strokefall genom att tillämpa dem på data från ögonspårning genom videoanalys. Målet och det avsedda resultatet är att potentiellt förbättra effektiviteten i prehospital vård. National Institutes of Health Stroke Scale (NIHSS) är en skala som traditionellt har används för att klassificera stroke. Genom att digitalisera NIHSS och använda den som en mall för att identifiera ögonrörelseavvikelser, en vanlig indikator på stroke, hoppas man uppnå detta. I denna studie, på grund av bristen på patientdata, skapades en hybrid datamängd innehållande både verklig och syntetiska data. De verkliga videorna (n=99) bestod av friska individer som simulerade blickförlamning i enlighet med NIHSS-riktlinjerna. De syntetiska datan var nödvändig eftersom vissa ögonrörelser, som när en patient har problem med kranialnerv tre, är mycket svår att härma. Därför användes digitala animationsverktyg (Blender och After Effects) för att skapa videor med syntetiska ansiktet (n=65) som simulerar dessa symtom. Den riktiga datamängden validerades sedan av en strokespecialist. De algoritmer som valdes för att tolka de kombinerade datamängderna var convolutional neural networks (CNN), deep neural networks (DNN), gated recurrent units (GRU), support vector machines (SVM) and long short-term memory networks (LSTM). En hybrid datamängd utökade mängden träningsdata, en avgörande faktor för att förbättra tillförlitligheten hos alla maskininlärningsmodeller. LSTM uppnådde det bästa övergripande resultatet i studien och visade en noggrannhet på 88%, en känslighet på 87,7%, en specificitet på 94,1% och ett F1-värde på 86,7%, vilket understryker dess framtida potential som ett tillförlitligt diagnostiskt verktyg i prehospital miljö. Sammanfattningsvis visar resultaten att tillämpningen av maskininlärning och videoanalys för att digitalisera och klassificera strokeinducerade ögonrörelser erbjuder betydande fördelar. Denna teknik har potential att förändra och fungera som ett effektivt komplement till traditionella metoder för strokebedömning. Innan dessa tekniker kan implementeras i praktiken krävs dock ytterligare forskning och förfining av metoderna.
Post
Exploring the Viability of Using WtE Incineration Ash as a Cement Replacement in Concrete
Eriksson, Joel; Chalmers tekniska högskola / Institutionen för arkitektur och samhällsbyggnadsteknik (ACE); Chalmers tekniska högskola / Institutionen för arkitektur och samhällsbyggnadsteknik (ACE); Baba Ahmadi, Arezou; Karlfeldt Fedje, Karin
The increasing importance of reducing greenhouse gas emissions demands new environmentally sustainable building materials. In the concrete industry most of the emissions comes from the production of cement. Research on materials that can be used as replacements for cement (SCMs) could help reduce the climate impact of the concrete industry. One of the possible SCMs that could be used are ash from WtE incineration. In this report the viability of the use of three different ash fractions from WtE incineration as cement replacement are be evaluated. These are a fly ash sample, an incinerator bottom ash sample and a mineral fraction incinerator bottom ash sample. Chemical and mechanical properties was evaluated using different techniques including a modified R3 method, compressive strength tests, PXRD and leaching tests with ICP-MS. It was found that all three of the ash fractions had chemical reactivity and the mechanical strength tests suggested that two of the ash fractions fly ash and MIBA had effect on the mechanical strength while the slag did not have ant effect. The results from the leaching tests were compared to thresholds on allowed limits of leaching from monolithic concrete samples and found to be below these thresholds. However, the limited knowledge and regulations on monolithic samples made the interpretation of these results difficult. The result from this thesis suggests that the use of WtE incineration ash a SCM is a possibility but that more research is needed. It is also concluded that the need for more and less complicated to understand regulations regarding leaching from monolithic concrete is needed in Sweden and that the current lack of regulations might be hindering the adoption of new sustainable building materials.
Post
Establishing photovoltaics in Sweden - A critical analysis of land use and win-win scenarios2024
(2024) Belzons Berthelemot, Oskar; Chalmers tekniska högskola / Institutionen för rymd-, geo- och miljövetenskap; Chalmers University of Technology / Department of Space, Earth and Environment; Berndes, Göran; Cederberg, Christel