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

RAIN AS RESOURCE: Rainwater Harvesting and Social Interaction
(2026) Jansson, Ida; Klang, Agnes; Rosén, Victoria; Wahlsten, Hugo
This project was carried out in collaboration with Melusi Youth Development Organisation, MYDO, a youth center located in the informal settlement Melusi in Pretoria, South Africa. The area is shaped by significant socio-economic inequalities and limited access to essential services such as sanitation, electricity, and clean water (IIED, 2022). MYDO provides both a safe and supportive study environment for children, as well as training programmes for youths to strengthen their skills and improve future employment opportunities (MYDO, 2026). The aim was to co-create, together with MYDO, a scalable and adaptable intervention that addresses water management challenges while improving social spaces. Through a frugal design approach, emphasizing locally resourced materials, the prototype shows how simple and effective solutions can inspire similar initiatives in other contexts. Codesign and co-building methods played an essential role throughout the project, with workshops and meetings strongly shaping both the design process and the final outcome. The water harvesting bench is a modular structure designed to serve multiple functions, including rainwater harvesting, social interaction, and gardening. Constructed from reused materials sourced on site, the bench can be adapted and replicated in different scales and variants.
Explainable AI for Network Intrusion Detection in Modern In-Vehicle Networks - An explainability pipeline for deep learning based intrusion detection systems for in-vehicle networks
(2026) Tian, Wenjun; Wang, Yexiao
Deep learning–based network intrusion detection systems (NIDS) have been widely adopted for detecting attacks in Controller Area Network (CAN) traffic due to their superior performance over traditional approaches. However, their black-box nature makes the underlying decision-making process difficult to interpret, limiting their suitability for safety-critical automotive environments. Existing studies on CAN NIDS have largely focused on detection performance and model design; limited work has examined shortcut learning through explanations or used explanations to improve the models. This thesis develops an explainable artificial intelligence (XAI) pipeline for DL-based CAN NIDS. Raw CAN fields are combined with explicit temporal and statistical features. Multilayer perceptron (MLP) classifiers are evaluated on the CAN-MIRGU and can-train-and-test datasets, and the autoencoder is evaluated on the CAN-MIRGU dataset. SHAP, LIME, Integrated Gradients, Trustee, and AE-p-values are used for behavior analysis, while Right for the Right Reasons (RRR) supervision and Gini-based attribution priors are applied to guide training. The MLP achieves high F1-scores on CAN-MIRGU, but its performance decreases substantially under unseen attacks and cross-vehicle tests in can-train and-test. Explanations show both meaningful use of timing-related features and reliance on dataset-specific payload patterns. For the autoencoder, p-value representations improve attack clustering over raw inputs. Moreover, p-values reveal that in the CAN-MIRGU dataset, DoS and fuzzing attacks are distinguishable, whereas spoofing and replay attacks remain difficult to separate. RRR increases the F1-score for non-zero-payload DoS attacks under distribution shift from 0 to 0.61. Gini-based regularization increases explanation sparsity while maintaining high predictive performance. These results show that explanation analysis is necessary for identifying shortcuts, assessing generalization, and developing more reliable CAN NIDS.
Evaluating LLM Adaptation Strategies in Physical Security Operations - Reducing Cognitive Load and Alarm Fatigue in Control Rooms
(2026) Lindberg, Olof; Christiansson, Casper
Physical security control rooms demand rapid decisions under high alarm volumes, producing cognitive load and alarm fatigue that degrade operator performance. This thesis evaluates three LLM adaptation strategies (zero-shot, few-shot, and retrieval-augmented generation) against a rule-based baseline on two tasks: alarm dispatch and guard chat. Evaluation was conducted in a simulation environment modelling GuardTools Command and Control, a SaaS platform for manned guarding operations. For dispatch, the rule-based strategy was fastest and cheapest. Among LLM strategies, few-shot achieved the lowest execution failure rate (1.82%), while zero-shot degraded sharply under clustered alarms, reaching 26.62% failures. RAG consumed roughly twice the tokens of few-shot without improving reliability. For chat, the decisive factor was context management rather than prompting technique: filtered few-shot and RAG isolated the relevant location description before prompting, achieving the highest quality scores while minimizing token usage.
Emission Scenarios for a 100% Renewable Faroese Power System A System-Level Life Cycle Carbon Intensity Assessment Towards 2040
(2026) Hedlund, Jakob; Ljungqvist, Konrad
The Faroe Islands have set an ambitious target of transitioning towards a fully renewable electricity system, but the climate impact of different technology pathways remains uncertain. This thesis assesses how alternative energy system configurations influence the system-level life cycle carbon intensity of the Faroese power system in 2040. Five scenarios were evaluated using outputs from a Python for Power System Analysis (PyPSA) based energy system model: a Base scenario, a Tidal scenario, an Offshore wind scenario, a Vehicle-to-Grid scenario and a Combined scenario including all investigated technologies. For each scenario, the system-level carbon intensity was calculated by combining technology-specific life cycle GHG emission factors with modelled installed capacity and annual electricity generation. Storage infrastructure, including battery energy storage systems and Vehicle-to-Grid (V2G), was included as a separate system-level contribution. The results show that the climate impact of a 100% renewable power system depends strongly on the available technology mix. Among the CO2-constrained scenarios, the Combined scenario achieved the lowest system-level carbon intensity, at 11.0 g CO2- eq/kWh, followed by the Tidal scenario at 33.8 g CO2-eq/kWh. The Base, Offshore wind and V2G scenarios showed much higher carbon intensities, ranging from 56.6 to 57.5 g CO2-eq/kWh. A key reason for this difference is the amount of battery storage required to balance the variable renewable generation. The Combined scenario required significantly less stationary battery storage, resulting in a much lower storage-related climate impact. The findings imply that achieving a low-carbon renewable electricity system is not only a matter of replacing the fossil generation, but also of integrating complementary technologies that reduce storage requirements, curtailment and capacity overbuilding. Tidal power appears especially valuable in the Faroese context due to its predictable generation profile, while V2G mainly contributes flexibility rather than direct emission reductions. Overall, the results highlight the importance of whole-system planning when assessing renewable energy transitions in isolated power systems.
Tournament on Path: Relation Abstraction and Ranking for Knowledge Graph Reasoning
(2026) Xiang, Shihao; Wu, Yihan
While Knowledge Graph Question Answering (KGQA) leverages Large Language Models (LLMs) toperform complex multi-hop reasoning, existing training-free frame works like Think-on-Graph (ToG) suffer from inherent limitations, including locally scoped pruning, unstable pointwise scoring, and the indiscriminate discarding of valuable candidates through random sampling. To address these challenges, this thesis introduces Tournament on Path (ToP), a novel reasoning framework that enhances relation abstraction and candidate ranking. The proposed framework systematically tackles the baseline’s flaws by implementing an entity-agnostic relation path pruning strategy that captures global path semantics and reduces the search space. To effectively operationalize this, ToP is instantiated into two specific variants. System 1 employs a hybrid lexical-semantic pre-filtering combined with a chunked tournament selection algorithm to stabilize ranking across large candidate pools. System 2 relies on purely semantic pre-filtering and a pairwise tournament selection method, introducing a topic entity masking mechanism to strictly prevent LLMs from answering using unverified internal knowledge. Experimental evaluations across four diverse benchmarks (CWQ, WebQSP, WebQuestions, and GrailQA) demonstrate that both variants consistently outperform the ToG baseline. The results confirm that these strategies not only improve reasoning accuracy but also resolve the "negative scaling" behavior, establishing a computationally efficient and consistently reliable framework for different language models.