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.

Visar 1 - 2 av 2

Senast publicerade

Investigating AI-Ethics through Necromancy - An Interdisciplinary Exploration of Uncertainties in Performative Technologies as Curated AI Personas
(2026) Hagman, Max; Kalm, Sally
This thesis was driven by its uncertain nature, where emerging AI technologies was investigated through Research through Design, Speculative Design and an artistic process. This perspective gave way to exploring solutions that would otherwise have been overlooked as unconventional. The project resulted in an interactive and performative art show, where a debate between four deceased media-philosophers was simulated through the use of Large Language Models. The show’s aesthetics, atmosphere and content all contributed to the User Experience. It was developed iteratively with formative and summative evaluations in form of interviews and surveys, with continuous on-site testing. The critical reflections on AI use-cases resulted in several ethical insights: there is a need to design for responsibility and accountability of generated outcome; there is power in orchestration, as the persons imitated with AI gain importance; there is a need for transparency in training data and agenda; and finally, anthropomorphism contributes to building relationships between humans and machines. Moreover, this thesis also reflects on the possibility of generating new knowledge, and concludes that personified AIs might be a tool to gain lost perspectives on current world states, but will not necessarily generate new knowledge with the current technological limitation. Lastly, This project calls for further research into charismatic AIs and guidelines for social chat-bots, as well as further research into integrity and consent when it comes to simulating people through AI technology.
Beyond Rides: Designing Internal Digital Systems for Dynamic Organization - Creating Guidelines for a Graphical Interface in Theme Park Organizations
(2026) Blomberg, Josefine; Vorberg, Amanda
The rapid digitalization of todays society, along with the emerging technologies it entails, places increasing demands on organizations and companies to adopt and implement digital systems. There are generally various systems being used simultaneously, and in order to remain effective and competitive, these need to have a high degree of user-friendliness in terms of being accessible, intuitive, and easy to use. This is of particular importance when it comes to organizations with high staff turnover, which is often the case in specifically service-based entertainment and leisure businesses. Among these are amusement park organizations, which typically rely heavily on seasonal employment, resulting in a constant flow and onboarding of new employees. Despite extensive research on usability in design and user experience (UX), there is a lack of studies regarding these aspects in relation to high employee turnover rates. This thesis project was carried out in collaboration with the amusement park Liseberg as a stakeholder, and the digital applications Gästis and Opening Hours as parts of their internal systems. The aim was to investigate what type of design approaches and practices best support usability in user interfaces (UIs) within organizations experiencing frequent employee turnover. The exploration was done by following the Double Diamond design process, with a total of three iterations. A user-centered approach was adopted, which involved semi-structured interviews and user tests conducted with employees at Liseberg, who regularly use any of the applications as part of their work role. The primary outcome of this study is the contribution to user-friendly UIs designed for use in high turnover organizational environments. The project resulted in the development of two high-fidelity prototypes in Figma, called Gästis 2.0 and Opening Hours 2.0, alongside a set of formulated design guidelines that can serve as a foundation for future investigation and improvement. The most important design guidelines identified are ensuring consistency across the interface, supporting efficient information scanning, providing clear and immediate system feedback, and following a structured and iterative design process. However, further research is recommended in order to validate feasibility and applicability of these guidelines beyond the specific context of this thesis.
Captioning Engine for AD/ADAS Data using Multi-Modal Large Language Models
(2026) Müntzing, Marcus; Johnsson, Noah
Autonomous Driving (AD) and Advanced Driver Assistance Systems (ADAS) are dependent on perception models that can understand and interpret their environment and surroundings. These models require a large amount of annotated data to achieve this goal. However, it is often difficult to curate large datasets as it is time-consuming, most footage is often redundant, and it is hard to find the rare edge cases that matter the most. Vision-language models (VLMs) offer a scalable alternative to automate the process of generating captions for driving scenarios. This thesis investigates how such captions can be used through representation learning for downstream tasks. First, a captioning engine combines rule-based detection of driving maneuvers (lane changes, cut-ins and cut-outs) with a VLM to generate detailed captions describing the environment and interactions in the scene. The rule-based component compensates for the VLM’s limited ability to reason about temporal progression across frames, while the VLM contributes rich descriptions of the surrounding scene. Secondly, the resulting video and caption pairs are used to fine-tune two contrastive vision-language embedding models, CLIP (ViT-L/14) and the Perception Encoder (L14-336), with the goal to align the captions and videos in a shared representation space. To our knowledge, this is the first use of the Perception Encoder in the AD/ADAS domain. Our results show that a rule-based classifier on dash-cam sequences could be effectively used as context enrichment of the caption generating VLM. Fine-tuning on these detailed and rich captions increases text-to-video Recall@1 from 25% to 83%. CLIP outperforms the Perception Encoder as the choice of backbone across all metrics and all maneuver types after fine-tuning. Further testing is needed to evaluate the generalization of this method beyond the maneuvers and dataset considered here. Future work includes extending the rule based system to additional maneuvers, such as lead-vehicle braking and vulnerable road-user interactions, and evaluating whether the learned representations transfer to other datasets.
Extrinsic Camera Calibration Using Vehicular Interior Features - A Domain-Agnostic and Targetless Pipeline
(2026) Mörck, David; Särnholm, Andreas
Advanced driver assistance and safety systems increasingly rely on interior sensing systems, such as Occupant Monitoring Systems (OMS) and Driver Monitoring Sys-tems (DMS). These systems depend on accurately calibrated interior cameras to provide reliable detection of behavior and positioning of the driver and occupants. Traditional camera calibration often relies on dedicated targets such as checker-boards, but such methods can be difficult to integrate into large-scale production environments. This thesis investigates targetless extrinsic calibration of an OMS camera by esti-mating its six degree of freedom pose from a single image of the car interior. Instead of using calibration targets, the proposed approach uses interior features as reference points. Two methods are evaluated: a relative method using a nominal reference im-age, and an absolute method using triangulated 3D points from known established views. The results show that extrinsic calibration using interior features is possible, with especially strong rotation estimation. Both methods presented here achieve errors as low as 0.09 degrees on synthetic data, while translation proved more challenging within the small mounting tolerances. The Absolute Method performed best for translation on synthetic data, reducing the error from an average of 5 mm to 1 mm, whereas the Relative Method showed no reduction in error. The Relative Method did however show better performance on limited real vehicle data by consistently reducing the reprojection error compared with the uncalibrated baseline. Overall, the thesis demonstrates the potential of targetless calibration for scalable OMS camera deployment. Although it is demonstrated on car interiors, the proposed pipeline is domain-agnostic and may be applicable to other environments containing repeatable visual features.
LiDAR Point Cloud Compression for Visualization - Utilizing compression techniques to enable the visualization of a LiDAR point cloud dataset
(2026) Bengtsson, Noa; Alowersson, Sofia
Methods to downsample and compress a large LiDAR point cloud dataset were developed for the purpose of enabling the visualization of a multi-minute sequence in a 3-dimensional OpenGL environment. The reduction in data is the result of segmenting dynamic objects and downsampling the LiDAR point clouds either by sampling a number of points in each cell in a 3-dimensional grid, or by computing an average point for each cell in a finer grid or octree structure. The ground plane can also be downsampled separately, with a different resolution, to give visual distinction between ground and environment. Together with a point cloud compression algorithm the final file size of a point cloud that is usable for visualization purposes can be reduced to 0.06%-3.6% of the original file size depending on chosen resolution.