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 publicerade
- Defect-aware 3D Geometry Reconstruction from 2D Annotations and 3D Scan Data(2026) Halwai, Abhishek; Ponkshe, DurveshComponents produced through additive manufacturing require particular methods of inspection to locate and characterize surface defects on those components. Although 2D image-based defect detection is an accepted method, linking that 2D defect information to the 3D geometry of the part by reconstructing the location and severity of the defect-induced deformation of the 3D surface is challenging with limited 3D scan data and the absence of certified 3D reference geometry. This gap between 2D defect annotation and 3D defect geometry motivates this thesis. This master’s thesis develops and analyzes a pipeline that projects 2D defect annotations onto a nominal 3D CAD model to reconstruct defect-aware 3D geometry with the help of structured-light scan data from the 3D-ADAM dataset. The pipeline projects 2D defect masks to 3D nominal surface points through the sensor’s organized point cloud, registers each PLY scan to the nominal CAD, computes the signed deviation between defective and nominal scans, and finally reconstructs the defect on the CAD model in the form of a point cloud, a mesh, and a deviation map. The thesis also explores the possibility of predicting defect deviation directly on the nominal CAD using a PointNet++ model. A geometry-only formulation shows significant limitations in predicting deviation magnitude and sign as this information is unavailable in the nominal geometry. After additionally providing a coarse 2D heatmap of the deviation as input, the magnitude prediction error was reduced by around 28% relative to the geometry-only model and 36% relative to the class-average baseline, and the direction of deviation was also recovered. The accuracy of defect placement onto the 3D surface is bounded by single-view registration and partially unresolved mirror ambiguity. This thesis validates the scan-based defect mapping and reconstruction pipeline, and describes the information required to predict defect deviation. Furthermore, this thesis identifies RGB-based defect prediction to reduce reliance on the 3D scan at inference as the key direction for future research.
- Security-aware Scheduling of Mixed-Trust Tasks in Multiprocessor Real-Time Systems(2026) Sjöberg, JakobReal-time systems are a fundamental component of safety-critical cyber-physical systems, such as autonomous vehicles, robotics, and avionics, where missed deadlines can have severe consequences. As these systems become increasingly interconnected, security has emerged as a critical concern, yet enhancing security typically increases computational demands, risking deadline misses. While security mechanisms for uniprocessor real-time systems are well researched, globally scheduled fixed-priority multiprocessor systems have received comparatively little attention. This thesis addresses that gap by introducing the Mixed-Trust task model for globally scheduled fixed-priority multiprocessor systems to enhance security, along with a corresponding schedulability analysis. The proposed model distinguishes between trusted and untrusted tasks, and upon detection of a security threat, transitions the system to a higher security level by introducing new security enhancing tasks while suspending untrusted ones. Inspired by the work on mixed-criticality scheduling the model additionally supports changing periods for tasks across security levels with the objective to increase the surveillance of the system. Finally to mitigate timing-based attacks, a randomization algorithm extending a uniprocessor algorithm, known as TaskShuffler, to globally scheduled fixed-priority multiprocessor systems is presented, diversifying task execution order to increase unpredictability for a potential attacker. The proposed mechanisms are evaluated empirically: the schedulability analysis is assessed using synthetically generated task sets, revealing how the task model and the number of available processors interact in ways that may help the system designer in selecting the hardware platform. The randomization algorithm is evaluated through simulation, demonstrating a meaningful increase in execution order unpredictability.
- Study of real time aspects for motion-driven music generation for a quadruped robot(2026) Hu, Shimin; Yi, XinRecent advances in legged robotics have expanded their deployment in human-centered environments, where rich multi-modal interaction is essential. While visual and kinematic behaviors are widely studied, real-time auditory feedback directly driven by robot motion remains comparatively underexplored. Such motion-driven sonification requires bounded low-latency synchronization between physical gait events and audio generation; however, standard general-purpose operating systems and middleware such as ROS 2 may introduce non-deterministic scheduling jitter, delayed callbacks, and audible timing artifacts under computational load. This thesis proposes and evaluates an embedded audio-motion synchronization framework for the Unitree Go2 quadruped robot. The system captures high-frequency proprioceptive sensor data at 500 Hz, applies lightweight filtering for gait-event extraction, and maps detected motion features to musical parameters and triggered audio responses in real time. To improve temporal predictability, the software architecture combines a decoupled C++ sensing/audio pipeline with operating-systemlevel real-time scheduling using a PREEMPT_RT patched Linux kernel and SCHED_FIFO thread prioritization. Experimental evaluation under simulated autonomous workloads, including SLAMand navigation-like CPU stress, shows that real-time scheduling primarily improves worst-case timing behavior rather than average computational speed. Under stress, the proposed real-time configuration reduces sensor data loss from 30.16% to 8.30%. It also substantially mitigates long-tail callback execution jitter: while the standard Linux scheduler exhibits a worst-case execution inflation of 1451× with a maximum callback time of 2,875 μs, the real-time configuration bounds the corresponding maximum to 24.7 μs, a 116× reduction. On the deployed Jetson platform, immediate footstep-to-audio events remain below 40 ms, well within the perceptual synchrony target used in this work. These results demonstrate that careful architectural decoupling and real-time scheduling can make motion-synchronous audio feedback practical for embedded interactive robotic systems.
- Optimization of battery AI models for edge deployment: A systematic study of model compression techniques and hardware-aware efficiency(2026) Xia, Yuxin; Liu, YeruiBattery management systems (BMS) increasingly employ data-driven models to estimate battery state of health (SoH), but deploying such models on automotive microcontrollers remains challenging because of limited memory and computational resources. This thesis evaluates a hardware-aware workflow for transferring, compressing, and preparing a long short-term memory (LSTM)-based SoH estimation model for embedded execution. The baseline model and its original training were provided by Cognivity AI; model development is therefore outside the scope of this thesis. The study uses the provided model and its established preprocessing pipeline, with battery-level data separation for evaluation. The workflow covers PyTorch to MATLAB model transfer, numerical equivalence verification, pruning and quantization evaluation, Simulink integration, and embedded C++ code generation. Two complementary compression directions are evaluated. Structured pruning reduces the number of learnable parameters while retaining useful predictive accuracy in the reported experiment. INT8 post-training quantization (PTQ) and quantization-aware training (QAT) reduce the theoretical weight storage to approximately one quarter of the FP32 value; QAT performs better than PTQ for the unpruned baseline in the reported comparison. Compression configurations are selected with validation data and evaluated on the held-out test data only after they are frozen. The selected models are exported to Simulink for model-in-the-loop functional checks, and statically allocated C++ code is generated using Embedded Coder. The generated artifacts were cross-compiled for the Infineon KIT_A3G_TC4D7_LITE evaluation platform. These results demonstrate conversion and compilation feasibility, but do not establish INT8 arithmetic in the generated implementation, on-target execution time, power consumption, or completed hardware deployment. Those measurements remain future work.
- Utilisation of Ferrochrome Slag for Partial Cement Replacement Technical Evaluation and Life Cycle Assessment of Water-Based and Alkali-Activated Systems(2026) Basaran, SibelThe high greenhouse gas emissions associated with cement production, together with increasing industrial residue generation, create a need for alternative binder materials. Ferrochrome slag (FCS) is generated in large quantities during high carbon ferrochrome production but currently exhibits limited utilisation potential due to low reactivity and chromium-related concerns. This study therefore evaluated processed ferrochrome slag (PFCS) as a partial cement replacement in water-based and alkali-activated cementitious systems. To improve the utilisation potential of FCS, PFCS was produced through slag treatment, water granulation, drying, and grinding to increase the amorphous phase content and reduce chromium-containing phases. Binder systems containing 10– 30% PFCS were initially screened using Vicat setting time measurements. Based on the screening results, systems containing 20% PFCS were further evaluated using isothermal calorimetry, BET, SEM, compressive strength, and Cr(VI) leaching. To evaluate the environmental performance of the selected mortars, a life cycle assessment (LCA) was performed. The experimental results showed that the 20% PFCS + Water systems exhibited hydration development similar to the cement reference despite prolonged setting behaviour, indicating participation of PFCS in the hydration reactions. Alkali activation further increased PFCS participation and resulted in higher cumulative heat release, although delayed hydration development was observed. Results from the environmental assessment further showed that PFCS incorporation reduced global warming potential compared with the cement reference. The water-based PFCS system showed the most favourable balance between hydration behaviour, chromium stability, and environmental performance. However, PFCS processing increased freshwater ecotoxicity impacts as a result of FeSi consumption during slag treatment. Alkali activation additionally increased chromium release together with surplus ore potential and freshwater ecotoxicity impacts associated with alkali activator and PFCS production. Overall, the results indicate that PFCS shows potential for partial cement replacement, although optimisation of processing and activation conditions is required to balance technical performance, chromium stability, and environmental impacts.