Electrical Engineering

Title: Sparse sampling for massive MIMO downlink CSI reference signals
Date 26 maj, kl 10-11 Place:: ES52 Student: Andreas Führ Examiner: Erik Ström
Abstract: The number of antennas used in massive Multiple input multiple output (MIMO) systems is expected to increase significantly to meet requirements of future radio access networks (RANs). In legacy 5G New Radio (NR), scaling the number of antennas imposes a proportional increase in overhead associated with the acquisition of channel state information (CSI) through downlink transmission of reference signals (CSI-RS). This thesis investigates methods to reduce the overhead of CSI-RS &nbsp;transmission using a twofold approach: optimising pilot placement for sparse sounding of the reference signals, and reconstructing the full channel information from these sparse measurements at the user equipment (UE).
We first formulate the sparse sampling of CSI-RS as a submodular optimisation problem, defining a cost function based on the frame potential of the downlink channel. A greedy algorithm for solving the sparse pilot placement problem is proposed and evaluated for a simulated 3GPP MIMO Urban microcell environment with a uniform planar array (UPA), achieving near-optimal pilot placement for subsets of antenna ports.
The second part of the thesis investigates the recovery of the full channel information from the sparsely sounded CSI-RS using a neural network (NN). We propose a physics-informed U-Net architecture that leverages the sparse angular representation of the downlink channel to recover the full channel information. The NN is trained on a large dataset of simulated noiseless downlink channels for the same 3GPP environment and for several different spatial pilot configurations and muting levels
The results show that the NN model can achieve a significant reduction in overhead while maintaining high reconstruction accuracy with comparable reconstruction performance to Basis pursuit denoising (BPDN), while outperforming BPDN in computational efficiency. The combined approach of optimising sparse pilot placement and using a neural network for reconstruction demonstrates the potential of machine learning techniques for improving the performance of massive MIMO systems in 6G networks and beyond.
&nbsp;Welcome!
Andreas and Erik

Title: H2 Engine Air-path and Lambda Control - Developing Adaptive Control Strategies for Optimal Combustion in Hydrogen-Powered Trucks
Date 26 maj, kl 14:00-15:00 Place: EDIT-rummet, Room 3364  Students: Gideon Geelnard &amp; Nandana Suryadevara Examiner: Jonas Fredriksson
Abstract:
Hydrogen internal combustion engines (H2-ICEs) offer a promising zero-carbon option to conventional diesel engines in the heavy duty industry. This technology present challenges in air-fuel ratio (AFR) control due to the wide flammability range and fast combustion characteristics of hydrogen. Maintaining the relative AFR ($\lambda$) value within acceptable limits is critical to reducing $NO_x$ emissions while ensuring stable engine operation. This thesis studies the feasibility of implementing an air-path $\lambda$-control strategy to improve AFR regulation in the engine. A grey-box modeling approach was used to map system parameters for the gas transport dynamics from engine test cell data during different operating conditions. The estimated parameters were then used to design and tune a feedback controller with dead time compensation. The control strategy was tested in a simulation environment, demonstrating improved $\lambda$ tracking with minimal impact on the engine performance compared to a baseline strategy. The results show that air-path lambda control is a promising method for managing the AFR in H2-ICEs, with potential for further improvements under extended operating conditions.
Welcome!
Gideon Geelnard &amp; Nandana Suryadevara

Welcome to the master thesis presentation
Design and Optimization of a Subsea Drive System with a Long Step-out
By&nbsp;Anton Apelberg and Tom Ekman
Master Thesis conducted at Onesubsea, under the supervision of Dimitrios Stellas and Kristiansen Baricuatro
Monday, May 26, 15:00 Fredrik Lamms room, Chalmers Electric Power Engineering division
Opponent: Juan Pablo Gomez and Jaime Pardo as our opponents Examiner: Peiyuan Chen
Abstract Subsea power transmission systems are essential for enabling offshore oil and gas operations, particularly as exploration extends into ultra-deep waters requiring long-distance electrical supply. The increasing demand for reliable and efficient subsea drive systems over extended cable lengths introduces several electrical and control challenges, including voltage drop, power losses, resonance effects, and harmonic distortion. This thesis develops a comprehensive model of a 150 km subsea AC transmission system using DIgSILENT PowerFactory to analyze steady-state voltage stability, power quality, and dynamic system behavior. The results demonstrate that proper cable sizing and reactive power compensation are crucial for maintaining voltage regulation along the transmission path. An open-loop scalar control strategy is applied to drive an induction motor, successfully achieving the desired steady-state speeds. However, transient oscillations observed during step frequency changes indicate the need for damping or filtering to improve dynamic response. Resonance analysis reveals harmonic amplification near every seventh harmonic, highlighting the importance of implementing harmonic mitigation techniques. A sinusoidal pulse width modulation (SPWM) method was integrated into the inverter model in PowerFactory, resulting in balanced, sinusoidal phase currents at the motor terminals and demonstrating stable operation during steady-state. In conclusion, the study confirms that a 150 km subsea AC drive system is technically feasible when supported by appropriate component selection, control design, and harmonic filtering. The findings underscore the importance of coordinated system optimization in achieving reliable and stable long-distance subsea power delivery..
Welcome! Anton, Tom, Dimitrios, Kristiansen, and Peiyuan

Titel: Positron Emission Tomography image reconstruction with time-of-flight data using a SPLAT projector
Students: Richard Nunstedt &amp; Erik Linghammar, MPMED &amp; MPENM Time and place:26/5 kl 15:00 in PJ-salen Opponent: Marcus Hansen Examiner: Ida Häggström Abstract: This thesis presents the development and evaluation of a physics-based system model tailored for image reconstruction in a novel time-of-flight (TOF) positron emission tomography (PET) brain scanner. The scanner design features continuous, multilayered monolithic detectors, offering improved spatial resolution and sensitivity compared to conventional PET systems. To fully leverage these hardware advances, a custom system model called the SPLAT model was implemented. The SPLAT model integrates key resolution loss mechanisms, including non-collinearity, detector blur, and coincidence timing resolution (CTR), using a Gaussian framework. In addition, positron range and photon attenuation, both implemented in a form viable for 3D, were incorporated to enable more realistic reconstruction. A simulation tool for PET list-mode (LM) data was developed based on this model, enabling controlled evaluation of reconstruction performance. Two simulation strategies were implemented to discretize the continuous detector domain. The forward and backprojection operators were implemented with performance optimizations in C and integrated into an iterative maximum-likelihood expectation-maximization (ML-EM) reconstruction pipeline. Image quality was assessed using normalized RMSE metrics and analyzed in relation to the modeled point spread function.
Welcome!
Richard, Erik and Ida

Welcome to the master thesis presentation
Power System Frequency Measurements and Requirements:&nbsp; Assessing the Needs in the Nordic Power System
Evaluating Frequency Constraints and Refining Normal Frequency Range for the Future Nordic Power System
By Diyasha Chenali Hapuarachchi
Master Thesis conducted at DNV, under the supervision of Lucas Thomée and Lester Kalemba
Tuesday, May 27, 09:00 Fredrik Lamms room, Chalmers Electric Power Engineering division
Opponent: Amuthavalli Jothiram Examiner: Peiyuan Chen Abstract The Nordic power system is undergoing a major transition, with large-scale renewable energy development expected over the next decades. The increasing share of inverter-based generation, higher production variability, and reduced system inertia will affect the frequency stability of the future Nordic grid. The aim of this thesis is to study frequency measurement methods, frequency sensitive components, and expected frequency variability under normal operating conditions of the future Nordic power system.&nbsp; Higher renewable penetration, with inverter-based connections introduces new challenges such as increased resonance and frequency variability, which increases the need of novel frequency measurement approaches. This study discusses these needs and commercial products such as relays, PMUs and controllers, that use different frequency measurement methods. The review of existing frequency measurement methods highlights various drawbacks of accuracy, signal quality and hardware implementation. Sensitive power system components for frequency such as generators, turbines, transformers, load side equipment etc. have also been discussed with a focus on the normal operational limits. A frequency variability study was conducted for the future Nordic power system based on the PSSE Nordic 44 model. The original Nordic 44 model was updated to reflect the current Nordic system and two simulation cases were developed considering operational snapshots for a winter day and a summer day. The frequency dynamic behavior was validated using actual frequency measurements. Then, two future scenarios for summer and winter were developed and the future system frequency variability was analysed by applying time series load changes.&nbsp; The results show that the frequency deviation from 50Hz increases with a greater variation in load. Increasing deviation rates are especially observed below 50Hz. Also, the summer day scenario showed a higher frequency fluctuation beyond the normal operational frequency limits compared to the winter day scenario. Future scenarios also showed higher frequency ramping compared to the present system model due to reduced overall system inertia. The findings indicate that the future Nordic power system will experience greater normal frequency variability with greater deviations and frequency ramps. Understanding system changes, being aware of sensitive components, adapting operational standards, and implementing novel frequency measurements to accommodate these changes will be crucial to maintaining system stability.
Welcome! Diyasha, Lucas, Lester, and Peiyuan

Titel: <span style="font-size: 12.0pt; font-family: 'Aptos',sans-serif; mso-fareast-font-family: Aptos; mso-fareast-theme-font: minor-latin; mso-bidi-font-family: Aptos; color: black; mso-ansi-language: SV; mso-fareast-language: SV; mso-bidi-language: AR-SA;">Federated Clustering of Electric-Vehicle Battery Usage Patterns for Personalized State-of-Health Estimation
Students: Axel Lundkvist, MPCAS Time and place: 09:30–10:30 AM, May 27th
Opponent: Felix Waldshock, Fredrik Sitje, Oskar Andersson and Ludvig Fornstedt Examiner: Changfu Zou Abstract: <span style="font-size: 12.0pt; font-family: 'Aptos',sans-serif; mso-fareast-font-family: Aptos; mso-fareast-theme-font: minor-latin; mso-bidi-font-family: Aptos; color: black; mso-ansi-language: SV; mso-fareast-language: SV; mso-bidi-language: AR-SA;">Accurately estimating battery state-of-health (SoH) in electric vehicles is hindered by the heterogeneity of usage patterns and privacy concerns around data sharing. Centralized battery management systems often fail to account for individual driving behaviors and environmental variations, resulting in imprecise SoH predictions. This thesis proposes a privacy-preserving federated clustering framework that groups EV battery usage sessions and trains personalized SoH models without exchanging raw data. Cluster-specific updates are aggregated via Federated Averaging to balance global insights with local adaptation. The approach is evaluated on representative battery cycling data, and preliminary results indicate its potential for enabling personalized SoH estimation under varied operating conditions.
Welcome!
Alex and Changfu

Welcome to the master thesis presentation
Integration of Electric Vehicles into Local Flexibility Markets: A Case Study of Aggregator Participation in Swedish Local Flexibility Market
By Fredrik Berntsson and Anna Oskarsson Tuesday, May 27, 14:00
Fredrik Lamms room, Chalmers Electric Power Engineering division
Opponents: Hanna Johnsson and Filippo Da Corte Vecchino Examiner: David Steen Abstract: Electric vehicles (EVs) are becoming increasingly widespread, transforming energy systems and presenting new opportunities and challenges. Beyond transportation, EVs can provide valuable services to the grid. Local flexibility markets (LFMs) allow electricity consumers and producers in specific regions to offer products that support local grid operation. This thesis explores the potential for EV integration in LFMs by identifying and reviewing operating European markets. The review finds that aggregated EV fleets are technically capable of participating in all studied markets. Additionally, a retrospective case study simulates EV aggregator participation in the Swedish LFM Effekthandel Väst during January 2025. Using a mixed-integer linear programming (MILP) model, the study minimizes charging costs for a fleet of 588 EVs under different scenarios. Results show significant cost reductions with V2G, particularly through a long-term availability contract, which nearly eliminates costs. Without V2G, a capacity limitation contract offers the highest savings, reducing costs by up to 37%.
Welcome! Fredrik, Anna, David, Ali and Tuan

Welcome to the master thesis presentation
Title: Power and signal integrity of printed circuit boards, Predicting real-world performance of copper interconnections using simulations
By David Einarsson Tuesday, May 27, 16:00 Uno Lamms room, Chalmers Electric Power Engineering division
Opponents: David Prytz Arcombe and Lowe Blank Examiner: Torbjörn Thiringer Abstract Printed circuit board (PCB) design is a complex process with small tolerances. A single fault could cause a malfunction of the complete system, requiring a costly redesign to regain a functional product. Introducing simulation tools in the design process could help identify problems with the PCB early in the design, reducing the need for physical prototypes, thus saving cost and reducing the environmental impact. This thesis investigates how different layouts affect the performance of the PCB in three key areas, including thermal behavior, power integrity (PI) and signal integrity (SI). Different design options were evaluated using simulations, which were later compared to measurements on physical boards. The simulated temperature and voltage drop of the PCB matched well with the measured values. In comparison, the impedance simulations had greater deviations from the measured values, where the differences were greater for power delivery networks compared to transmission lines. However, the impedance simulations of alternative designs gave a good visualization of the differences between the designs, which agreed with the general differences captured by the measurements. Some scenarios, including close decoupling placement and far-end crosstalk, were harder for the simulation to accurately predict. While stripline transmission lines and near-end crosstalk were easier to simulate. Changing the design of copper interconnects in the layout played a big role in deciding the final performance of the PCB, concerning thermal behavior, PI and SI. Utilizing simulation software in the design process allows for a good comparison between design alternatives, without the need to order a prototype. However, exact values of a certain design could be harder for the simulation to accurately estimate.
 Welcome! David and Torbjörn

Title: Designing a Web-Based Graphical User Interface for Managing Digital Twin Simulations in Healthcare&nbsp;
Students: Alva Cider and Norea Ronnby Examiner: Stefan Candefjord Opponents: Wilma Bergman and Susanne On Huang Date/time: 28 May, 2025 10:00-11:00 Place: Landahlsrummet 7430
Abstract:&nbsp; With an aging population, a growing shortage of healthcare professionals, and increasing demand for in-home care solutions, the need for digitization in healthcare is becoming more urgent. Integrating multiple medical devices into large-scale systems adds complexity, making thorough testing essential prior to real-world deployment. At Chalmers University of Technology, there is an ongoing research initiative developing a digital sandbox environment for managing digital twin simulations in healthcare. This thesis aimed to design a web-based graphical user interface (GUI) for the sandbox environment. Using a human-centered design approach, the GUI was developed through an iterative process involving prototyping and user testing, grounded in a realistic fall detection simulation case. The resulting interface features two integrated but distinct views: a patient-centered view for clinical users and a component-oriented view for technical users, enabling effective cross-disciplinary collaboration. Although still an early prototype, the GUI provides a solid foundation for further development and serves as a foundation for future development and integration of more complex scenarios.
Welcome!
Alva, Norea and Stefan

Title:&nbsp;Design and Modeling of a Modular Payload Carrying Tethered Drone System
Students: Fredrik Nygren och Oscar Schyum
Time and place:&nbsp; 2025-05-28, 10:00 E2 Room 5430 Femman
Examiner: Martin Fabian
Supervisors:&nbsp; Niklas Eisele, SAAB Jens Grönberg, SAAB
Opponents: Maulik Rakesh Rajput Åke Sterning
Abstract: Unmanned Aerial Vehicles (UAVs) are currently employed in a plethora of applications, spanning both military and civilian sectors. A key limitation of conventional UAVs is their restricted flight time and operational range. Efforts to extend this range often come at the expense of reduced payload capacity. Tethered drones offer a promising solution to overcome these limitations, where power and communication are transmitted through a tether, allowing the system to achieve substantially longer flight times, compared to conventional drones.
This thesis defines the capabilities and limitations of a tethered drone, designed for high-capacity and modular payload applications. The system is constrained by predefined dimensions for the drone, ground base and the upper &amp; lower payload. The system is separated into different subsystems, each of which is investigated individually based on a series of research questions. Design decisions are based on the predefined constraints, along with additional requirements derived from a range of potential applications. Furthermore, a concept generation and evaluation phase is conducted, during which concepts are iteratively assessed through performance-based comparisons. A hexacopter with three-bladed foldable propellers, capable of landing in any yaw-angle, emerges as the most suitable solution. This concept offers additional margins in thrust compared to other viable solutions. This additional thrust can be allocated to increase acceleration, extend tether length, enhance payload capacity, or prolong untethered flight duration.
Following this, the hexacopter and tether are modeled using the Euler Lagrange method, with the tether modeled as a lumped mass model. The modeling, based on a comprehensive literature review, identifies the appropriate governing system equations and results in a coupled system of ADEs. The system is subsequently controlled using cascaded PID control and simulated in Simulink to validate the system design and to assess its performance in terms of stability, robustness and redundancy. Key performance metrics, including recommended tether tension, optimal payload distribution, and maximum operating time are also computed by means of simulation under various environmental conditions and payload variations.
Welcome!
Fredrik, Oscar and Martin

Title: Automatic Classification of Contaminants in Surface Electromyography Signals: Statistical Evaluation of Signal Features and Machine Learning Models for Real-Time Implementation&nbsp; Students:&nbsp;Annie Abrahamsson Eknefelt &amp; Ludvig Swala Supervisor: Morten B. Kristoffersen Examiner: Giacomo Valle
Time: 11:00 - 12:00 Place: Landahlsrummet 7430
Opponents: Elsa Threström &amp; Elvira Carlén
Abstract: Can the most prevalent contaminants in sEMG signals be accurately identified and classified during an sEMG examination? sEMG and its many applications are highly affected by signal contaminants, decreasing efficiency and accuracy. This thesis presents a machine learning-based method for automatically classifying sEMG contaminants. A comprehensive feature evaluation was conducted to identify signal features that could differentiate signals of different contaminations and signal-to-noise ratios (SNR). Combining time and frequency domain features improved the ability to distinguish between different contaminants at various SNR levels, enhancing overall classification accuracy. Publicly available sEMG signals were artificially contaminated to simulate real-world noise. Among several machine learning algorithms, a Random Forest model type achieved the highest accuracy. Two different models, one single-label model and one multi-label model, provide the possibility to either detect the most prominent contamination or all present contaminants. 5-second segments for classification allow classification to be performed during sEMG examinations and implemented in devices such as prostheses. The results suggest that the proposed approach enables robust identification of sEMG contaminants, allowing direct feedback of contamination during sEMG examinations, providing the opportunity to improve the quality of sEMG signals, improve diagnostic accuracy, and improve patient quality of life.&nbsp;
Welcome!
Annie, Ludvig and Giacomo

Title: Modeling and Control of Aeration in a Wastewater Treatment Plant Students: Nils Larsson and Daniel Svensson&nbsp; Examiner: Veronica Olesen
Time:&nbsp;28/5 kl. 13:15-14:00 Place: EL53
Abstract: In this thesis, several modified control strategies for aeration control in a wastewater treatment plant are implemented. Their performance with focus on energy consumption and airflow pattern are evaluated and compared to the existing control system. In the wastewater treatment process, the biological step is where the oxygen is needed to regulate the effluent ammonium levels, which can therefore be controlled with the airflow. A model of a WWTP was built in Simulink to simulate the plant and evaluate the different control strategies. Time periods for real world data were selected and used, both for evaluation of the model accuracy and the control system performance. The final conclusions made are that there is room for improvement of the control system with for example feedforward of the inflow rate. Improvements are however limited to the daily variations and cannot help suppress the very high spikes in ammonia concentration seen a few times every year. The likely cause being that these are present due to capacity constraints being reached in the system. Energy consumption was also not substantially lowered by the modified control strategies and therefore, the implementation by them is probably not sensible. Attempts were also made to implement a LQG controller but failed due to two main reasons. The first being the nonlinearities of the system being too large to use a linear model based controller, the unstable nodes of the model, making the system only locally stable.&nbsp;
Welcome!
Nils, Daniel and Veronica

Title: Self-supervised Pre-training with Vision Foundation Models on 3D Point Clouds Subtitle: Comparing Downstream 3D Object Detection Performance Across Varying Label Quantities for Two Pre-training Strategies
Time and place: 28/5, 16.00 and Room 3364&nbsp;
Opponents: Osama Alsheikh Ali and Sizhe Chen Examiner: Lars Hammarstrand
Abstract: The ability to accurately identify objects and their spacial properties in the 3D domain is a critical task within the field of autonomous driving. To achieve this goal several computer vision techniques can be utilized including employing deep neural networks that given a LiDAR point cloud can identify objects and their corresponding 3D bounding boxes. For deep models to learn general properties of the environment and be able to identify present objects in it large amounts of labelled data are required. Since labeled data is expensive and time-consuming to obtain, it is crucial to reduce reliance on it without compromising performance. Therefore, this thesis investigates two pre-training tasks for 3D object detection on LiDAR point clouds and highlights their effectiveness on one-shot and few-shot detection. The pre-training methods in this project utilises two different web-scaled foundation models, namely, SAM2 and DINOv2. As both models are trained on 2D images, we lift their outputs to 3D to generate pseudo-labels for training a 3D object detection model in a self-supervised fashion. Predicted features from DINOv2 are used in the first pre-training task, while segmentation masks from SAM2 are used in the second task. As a result, the performance of each method reflects both the choice of pseudo-labeling method and the quality of the underlying foundation models. The results show that the model effectively learn to identify an object type from a single annotated sample after pre-training on either SAM2 or DINOv2. With only a few samples available, the pre-trained and fine-tuned models significantly outperform the baseline which is trained fully-supervised from scratch. This emphasises how these pre-training methods enabled the model to learn and recognize less common object categories even when they are represented by only a few instances in the dataset.&nbsp;
Welcome!
Pauline, Anni and Lars

Title: Adaptive Actuator Friction Modelling for Control Students: Johan Gustafsson &amp; Dennis Hjertén, MPSYS Examiner: Martin Fabian Opponents: Valdemar Samuelsson and George Vasanth Fransua Time: Friday, 30 May, 13:15 Place: E2 Room 2503 Fredrik Lamm
Abstract: Friction is a non-linear disturbance present in all electromechanical systems, which produces an opposing force or torque that resists motion. The effects become apparent and particularly problematic at low velocities, where the nonlinear friction characteristic reveals itself through phenomena such as limit cycles and stick-slip motion. This complex behavior presents significant challenges in high precision motion control systems. This thesis investigates methods for friction compensation in electromechanical actuators.
Compensation strategies are implemented on an Engine Control Unit (ECU) which controls an Electronic Throttle Control (ETC) unit in real-time. Experimental results demonstrate that friction compensation techniques can significantly enhance the tracking performance and accuracy of the ETC at low velocities. Furthermore, the thesis highlights the trade-off between control signal activity and tracking performance, showing that different methods of compensation exhibit different characteristics in how they balance these objectives.
Welcome!
Johan, Dennis and Martin

Title: Machine learning-based prediction model for in vitro fertilization success Students: Elvira Carlén and Elsa Thoreström&nbsp; Time and place: Monday 2/6 kl.10:00 in Analysen Opponents: Annie Abrahamsson Eknefelt and Ludvig Swala Examiner: Ida Häggström
Abstract: Patients undergoing in vitro fertilization (IVF) are often informed about their average chances of achieving a live birth, which is mainly influenced by the women’s age. This project is part of a larger study and aims to predict individualized chances of success using machine learning (ML), helping patients set realistic expectations and guide their decisions.&nbsp; Three research questions were addressed, each focused on predicting the likelihood of a live birth in one of the following scenarios: the second cycle following an unsuccessful first started cycle; the second or third cycle following an unsuccessful first cycle; and the third cycle following an unsuccessful first and second cycle. A complete IVF cycle includes ovarian stimulation, oocyte retrieval, fertilization, embryo culture and all resulting fresh and frozen embryo transfers. The dataset used to train the ML models was collected from Sahlgrenska University Hospital between January 2016 and December 2022 and included 21 features from 5,170 cycles. Three ML models were evaluated: logistic regression, random forest (RF) and support vector machines (SVM). All models achieved a higher area under the receiver operating characteristic (AUROC) curve when predicting the outcome after a failed first IVF cycle, compared to predictions after an unsuccessful first and second IVF cycle. The predictive performance of logistic regression and RF was consistent with results from similar studies that used datasets comprising more than 70,000 IVF cycles. This study suggests that comparable results can be obtained with a smaller dataset and with fewer adjustments, such as accounting for non-linear relationships between the features and the live birth outcome in logistic regression, which RF handles naturally.
Welcome!
Elvira, Elsa and Ida

Title: Circular package delivery with a fixed-wing drone
Students: Anton Carlsson, Hugo Rådegård, MPCAS Supervisor: Fredrik Falkman Examiner: Balazs Kulcsar Place and date: E2 Room 3364 EDIT-rummet, 2025-06-02 11:00
Abstract: The goal of the project was to investigate the possibility of circular package delivery using a fixed-wing drone on behalf of SSRS. The main tool used to complete the project was simulation using Matlab and the Matlab toolboxes of Simulink and UAV toolbox. Three programs were created: One to study free fall and the factors affecting the drop, another was created for the circular drop as well as a program for the baseline fly-by drop to compare the results with. In an attempt to improve the dropping accuracy, a disturbance function was created based on the error of a turbulence free simulation of a singular drop. The results of the project was that high accuracy (sub 1 m error) can be achieved using the circular method at a higher and more reliable rate than the baseline fly-by drop, at a severe cost of time. Further studies that apply what was learned during the simulation in practice are needed before labeling the circular package delivery as a viable method to be used by SSRS.
Welcome!
Anton, Hugo and Balazs

Title: Adversarial Inverse Reinforcement Learning for Energy-Efficient Marine Vessel Operation
Supervisor: Simon Johansson, Cetasol AB Names: Anish Janardhan, Muddur Ajay Nayak, MPSYS Examiner: Balazs Kulcsar
Place and date: E2 Room 3364 EDIT-rummet, 2025-06-02 13:00
Abstract: The maritime industry serves as a backbone of global trade, transporting nearly 80% of goods worldwide, making it indispensable for the global economy. Yet, this sector remains a notable contributor to greenhouse gas (GHG) emissions, accounting for 2-3% of total emissions-a figure projected to rise dramatically by 2050 under business-as-usual scenarios. Addressing this escalating environmental impact requires innovative strategies for optimizing vessel operations and minimizing fuel consumption, which are crucial for sustainable maritime practices. This thesis focuses exclusively on the application of adversarial inverse reinforcement learning (AIRL) to tackle these challenges. AIRL leverages expert demonstrations from seasoned captains who have successfully navigated complex and diverse maritime conditions. By analyzing these expert behaviors, AIRL derives comprehensive reward function that encapsulate optimal operational strategies, paving the way for the development of adaptive control policy. The strength of AIRL lies in its ability to capture the nuances of expert judgment, enabling marine vessels to emulate the energy-efficient decisions made by experienced captains. Unlike traditional control systems, which often struggle with the nonlinear dynamics and diverse challenges of maritime environments, AIRL offers a robust data-driven solution tailored to the specific operational needs of vessels. By integrating AIRL, this research aims to enhance energy efficiency, reduce fuel consumption, and lower carbon emissions. In doing so, it contributes not only to the operational goals of modern marine vessels but also to the broader environmental objectives of the global community. This work aligns with the maritime industry’s transition toward greener practices and supports the adoption of sustainable solutions in the face of growing ecological concerns. Through the exclusive application of AIRL, this thesis underscores its potential to drive transformative change in the maritime
Welcome! Anish, Muddur and Balazs

Title: Developing a Heuristic Algorithm for Solving an Extended Version of the Fleet Size and Mix Vehicle Routing Problem
Names: Fatemeh Alibakhshi and Pawel Kasinski, MPENM and MPALG Academic Supervisor: Jiaming Wu, ACE Supervisors at Company: Sunney Fotedar and Toheed Ghandriz, Volvo Group
Examiner: Balazs Kulcsar Place and date: E2 Room 3364 EDIT-rummet, 2025-06-02 14:00
Abstract: As electric trucks are being commercialized, new challenges related to infrastructure, routing, and investment planning emerge. For instance, the problem of determining optimal routes for a fleet of vehicles that needs to deliver goods to customers cannot be solved as it has been done previously. In this master thesis, a heuristic algorithm is developed to solve an extended version of this problem, denoted the vehicle routing problem, that includes aspects such as a heterogeneous fleet, pickups and deliveries, charging, and split deliveries. The algorithm is based on the Adaptive Large Neighborhood Search framework, meaning that it finds approximate solutions, but has the advantage that it can solve realistic large-scale instances of the problem.
Welcome!
Fatemeh, Pawel and Balazs
&nbsp;

Title: Industrial Video Anomaly Detection using a a weakly supervised Predictive Autoencoder
Presenter: Petter Dittmer Examiner: Giuseppe Durisi Opponents: Shahir Imam and Elliot Källander
Date and venue: Monday, June 2 at 16:00, E2 Room 2503 Fredrik Lamm
Abstract: This thesis provides a predictive autoencoder model and data pipeline for detecting anomalies in production flow in industrial environments. It also contains a comparison between three commonly used modifications in deep learning models, being spatial attention blocks, dropout and skip connections. The project was done in cooperation with the company EyeAtProduction AB in Borås. The model is designed for flexibility, robustness and short training times in new environments, rather than state of the art performance. It uses a pretrained version of the image recognition network ResNet-18 for encoding sequences of four video frames. The encoded frames are merged with a (1 × 1 convolution operation, and then decoded via transposed convolutions, resulting in a prediction of the next frame following the sequence. This prediction is then compared to the actual next frame. By only training the network on footage of normal production, it will become proficient at predicting normal movement and spatial features, but struggle to reconstruct anomalous sequences and objects. By calculating the PSNR score between the prediction and the true frame, an anomaly score can be assigned to each incoming sequence during inference, and any sequence exceeding a set threshold can then be flagged as anomalous.
The model and its variations show promise in both controlled environments and real-world cases, but even with heavy data augmentation they are still sensitive to lighting changes and vibrations in the camera, making them prone to false positives. More research would need to be done to minimize this problem further, but possible solutions could be collecting larger and more diverse training sets, and making the threshold adapt to the long term shifts in the prediction scores during inference.
Welcome! Petter and Giuseppe

Title: Evaluation of ancillary services from wind turbines
Students: Nowroz Bin Nasim &amp; Kristoffer Olsson Examiner: Ola Carlson, Sara Fogelström
Date: 3 of June at 9:00 Room: Uno Lamm, 3571
Abstract: There is currently limited guidance on evaluating the performance of wind turbines in delivering ancillary services based on measured operational data. This thesis presents an evaluation method that uses grid frequency variations to estimate droop and ramp rate without relying on the predefined step signals used in Svenska kraftnät’s prequalification. The method is applied to a simulation of Chalmers wind turbine and to data measured at the commercial wind farm, focusing on FCR-D Down. Two days with different operating conditions were analysed using the same procedure. The study looks at power response, droop, ramp rate, and how well the response matched the accepted bids. The results show that the windfarm responded faster and with a stronger droop than the simulated turbine, which acts as a reference for Svk’s requirements. The method can support further evaluation of frequency control from wind farms and help wind farm owners choose turbines that are suitable for frequency control.
Most Welcome Nowroz Bin Nasim, Kristoffer Olsson, Sara Fogelström and Ola Carlson

Title: <span style="font-size: 12.0pt; font-family: 'Aptos',sans-serif; mso-fareast-font-family: Aptos; mso-fareast-theme-font: minor-latin; mso-bidi-font-family: Aptos; color: #242424; mso-ansi-language: SV; mso-fareast-language: SV; mso-bidi-language: AR-SA;">Transfer Learning for Battery Health Forecasting: From Lab to Real World Data
Students: Oskar Andersson and Ludvig Fornstedt
Opponent: Axel Lundkvist
Examiner: Changfu Zou
Date: June 3rd, 13:00-14:00 Room: E2 Room 5430 Femman
Abstract:
Accurately predicting battery health in electric vehicles presents a significant challenge due to the scarcity and variability of real-world data. While laboratory tests provide precise, controlled measurements, they often fail to capture the diverse and unpredictable conditions of actual vehicle usage. This thesis addresses this gap by developing a transfer learning framework for personalized battery degradation modeling, leveraging small amounts of real-world data to enhance the generalization of Remaining Useful Life (RUL) predictions. By integrating vehicle-specific patterns with generalized trends, this approach aims to bridge the lab-to-road divide creating adaptive battery management systems.
Most Welcome Oskar Andersson and Ludvig Fornstedt

Title: Synergies between Renewable Energy and Water Management Systems By Nathalie Pettersson Opponent: Tobias Filmberg&nbsp; Examiner: Jimmy Ehnberg&nbsp;
Location: Fredrik Lamm Date: June 4th at 08.30
Abstract: Water, energy and food systems are essential for human development and a modern life but face growing, complex and overlapping challenges under climate change. This thesis examines these issues in a case study in eastern Rwanda, focusing on irrigation and water purification under climate change. A linear optimization model was developed that sized and decided on operation for cost-effective electrified irrigation systems using two different set-ups. One, combining solar photovoltaics and biodiesel generators, and another one with only solar energy. Both met energy demand sustainably but the biodiesel system was more cost-effective and flexible. The results showed that slightly oversizing the system enabled water purification at lower cost than separate sizing of each system. Climate change modelling cases revealed cost increases of up to 6% for irrigation alone and 23% when combined with purification. This highlights the need for strategic resilience planning. Furthermore, prioritization strategies strongly influenced system performance when unadapted to climate change. The findings highlight that sustainable electrified irrigation is achievable. Although, integrated designs and careful trade-off navigation in the Water-Energy-Food Nexus remain critical under climate uncertainty. Broader challenges, including infrastructure, materials, and political factors, pose other pressures regarding these systems.&nbsp;
Welcome!
Nathalie and Jimmy

Title: A Viability Analysis of Multipurpose Dam System in Rwanda By Amelia Bergum &amp; Alida Nilsson, MPSES
Opponents: Cajsa Jacobsson &amp; Hugo Palmgren Examiner: Jimmy Ehnberg&nbsp;
Location: Fredrik Lamm Date: June 4th at 09.30
Abstract: In Rwanda, where agriculture accounts for 25% of the national GDP, improving productivity is essential due to limited opportunities to expand cultivated land. In parallel, expanding access to both electricity and clean drinking water remains a central development challenge. National strategies promote a diversified energy system, with off-grid solutions widely encouraged for rural electrification. This thesis investigates the feasibility of implementing a multipurpose dam system that addresses irrigation needs while simultaneously enabling electricity generation and/or water purification through integrated infrastructure. A linear optimisation model was developed to identify the most cost-effective and robust system configuration, dimensioned to meet the irrigation needs of the focus area. This base model was then extended to explore multifunctional solutions by integrating electricity generation and/or water purification in a series of scenarios. The results show that a multipurpose dam providing irrigation, electricity, and clean water can be a viable solution for Tabagwe. Depending on local needs and economic capacity, different&nbsp;service levels can be integrated. The most advanced configuration, delivering Tier 3 electricity and 50 litres of purified water per person per day, resulted in a total annual cost of $114 729. This setup included UV water purification, which was technically feasible without requiring system resizing. However, the high capital cost of the purification units made this solution financially demanding. At the same time, the system demonstrated substantial surplus energy, highlighting the&nbsp; potential for integrating alternative purification technologies with higher energy requirements but lower costs. The integration of battery storage was also identified as a promising option to increase system flexibility and improve energy utilisation.&nbsp; Welcome!
Amelia, Alida and Jimmy

Title: Transformer-Based Multi-Object Tracking of Football Players Using Pseudo-Labeling
Students: Anthon Odengard &amp; Anton Heden Date: 4/6 Time: &nbsp;9-10 &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;&nbsp; Place: 3364 EDIT-room Examiner: Tomas McKelvey
Abstract:&nbsp; This thesis investigates the application of transformer-based architectures for tracking football players in static-camera video sequences. To address the challenges of multi-object tracking (MOT) in complex football environments, we adapt the MOTRv2 transformer-based tracker specifically for this domain. Our methodology incorporates a pseudo-labeling framework that leverages unlabeled data, enabling semi-supervised training and progressive model refinement. Additionally, a re-identification strategy is implemented to enhance tracking consistency. We evaluate the adapted model on football footage using metrics focused on identity preservation and tracking accuracy. The results demonstrate the potential of transformer-based trackers in sports analytics and emphasize the advantages of semi-supervised learning in scenarios with limited annotated data.
Welcome!
Anthon, Anton and Tomas

Title: Replicating human scalp properties for enhanced accuracy in the testing of bone conduction hearing devices
Students: Anna Fassih &amp; Elina Sandoff Examiner: Sabine Reinfeldt Supervisor: Dan Nyström, Cochlear Bone Anchored Solutions AB Opponents: Tove Thedin Olsson &amp; Alexandra Johansson, MPBME
Date: Wednesday 4th June 09.00 Room: Landahlsrummet, 7430
Abstract: This thesis investigates and develops an artificial human scalp replica for use in bone conduction hearing device testing, with the goal of improving the accuracy and reliability of bone conduction hearing device evaluations. The work was motivated by Cochlear Bone Anchored Solution’s need to enhance their existing head simulator, which previously lacked a component mimicking the mechanical properties of human scalp tissue. To achieve this, a wide range of silicone-based scalp materials were evaluated using three primary performance metrics: mechanical point impedance, attenuation and acoustic feedback. All measurements were conducted using a standardized measurement setup involving Cochlear’s HeadSim1 simulator. Additionally, a market investigation was conducted to collect new mechanical point impedance reference data from 19 human subjects under identical conditions to ensure comparability with the artificial materials. Multiple types of artificial scalps, including Ecoflex and FS10 variants, in-house crafted silicones, and molded custom shapes, were assessed through repeated measurements to ensure consistency and minimize errors. The findings reveal that Ecoflex0010, particularly the adaption developed in this thesis, best replicates human scalp properties. It demonstrates the lowest mean absolute error (MAE) in mechanical point impedance and attenuation combined, and performs well within acceptable ranges for acoustic feedback compared to clinical reference data. The aim of this study was fully achieved: a validated, usable artificial scalp that improves the fidelity of bone conduction hearing device testing was identified. The broader implication is an enhancement in preclinical evaluation tools for bone conduction devices, leading to more reliable data and potentially reducing the need for clinical testing iterations. This contributes to better design decisions, increased patient safety, and faster innovation cycles in the field of bone conduction hearing device technology. 
Most Welcome! Anna Fassih &amp; Elina Sandoff

Titel: A Study on the Effect of Incorporating Spinal Radiographs into Multimodal Neural Networks for Osteoporotic Fracture Risk Assessment
 Date and time: 4 June, 10:30-11:30&nbsp;
 Room: E2 room 3364 EDIT-rummet Student: Erik Blomqvist
 Opponent: Viktor Lindfors, Cecilia Nordberg
 Examinator: Jennifer Alvén
 Abstract:
Abstract: Osteoporosis, characterized by reduced bone mass and micro-architectural deterioration, increases the risk of fractures, particularly in elderly populations. Traditional fracture risk prediction models rely on clinical risk factors but do not incorporate imaging data, potentially overlooking structural indicators. This thesis explores the potential of multimodal learning by integrating spinal X-ray images with clinical risk factors to improve fracture risk prediction. Both full spinal radiographs and cropped vertebral images are evaluated to investigate whether focusing on anatomically relevant regions can enhance the predictive signal. Two deep learning architectures are considered, Convolutional neural networks (CNNs) and Vision transformers (ViTs), alongside multiple fusion strategies for combining image and tabular data. Results demonstrate that multimodal models consistently outperform baselines, with the best performance achieved by a CNN using vertebral crops and intermediate fusion (C-index: 0.69, AUC: 0.76, Brier score:0.14). This suggests that image data alone contain meaningful predictive information and that combining imaging with clinical features enhances fracture risk prediction. Vertebral crops generally yielded better performance than full radiographs, highlighting the importance of localized features. However, the models were evaluated on a single dataset of elderly Caucasian women, indicating the need for future work to improve generalization across diverse populations.
&nbsp;
Welcome!
Erik Blomqvist

Titel: Multi Camera Extrinsic Calibration Undertitel: Online Refinement for the Rotations of Fisheye and Long Range Cameras Mounted on a Moving Car using Natural Feature Detection in Overlapping Fields of View
Student: August Blanc Opponenter: Eric Wahlström and Efraim Zetterqvist Examiner: Lars Hammarstrand Slutpresentation: Onsdag den 4e Juni 13:45-14:30 i rum 3364
Abstract: Autonomous driving has potential of increasing traffic safety and enabling greater automation in our society. These systems rely on high precision sensor data, not the least from cameras, in order to detect and map out the environment around a vehicle. In doing this, knowledge about how the sensors are mounted is of crucial importance. This is typically calibrated in a static rig upon manufacturing a vehicle, but after driving for a while the orientation of sensors might change due to bumps and vibrations, thus requiring regular recalibration.
This thesis presents an automated refinement method for the rotations of cameras mounted on a car. There are four fisheye cameras with a wide field of view, that together can be used to form a 360 degree view of the surrounding ground. This is typically shown on the display when parking a modern car. There are also three regular cameras, one looking straight forward and two looking backwards and slightly to the left and right, respectively. Using images taken by the cameras at the same time, we find pixel coordinates that correspond to the same real world point for cameras with pairwise overlapping fields of view. Using this information we then optimize all camera poses simultaneously, starting from an initial pose estimate.
The performance is evaluated by two criteria. Firstly, we consider the surround view images that are created by stitching together fisheye camera images projected onto the ground plane. Looking at the overlap between cameras, the alignment of e.g. lane markings and shadows in the ground plane indicates the quality of a calibration. Secondly, we annotate point correspondences for each overlapping camera pair in each dataset. These are used to calculate a quantitative error metric, essentially telling us how close the pixel coordinates are to actually originate from the same real world point for the given camera poses.
The results indicate that our method achieves an accuracy comparable to that of a static calibration rig. This is a very promising result, since the time and space consuming rig calibration could possibly be switched for a calibration method that can run while the car is driving. Additionally, the fisheye camera calibration appears to achieve an accuracy comparable to other state of the art methods, while also being applicable to a wide range of driving environments and allows for other cameras to be used in the calibration as well.
Welcome!
August and Lars

Titel: A multi-task learning approach for identifying white matter hyperintensities and vascular cognitive impairment in FLAIR MRI images
Date and time: 5 juni 10.30-11:30 Room: E2 room 3364 EDIT-rummet
Students: Amanda Hoveklint &amp; Kajsa Homann Opponents: Saga Frisell &amp; Kristoffer Gustavsson Examiner: Jennifer Alvén
Abstract: This thesis investigates the potential of using FLAIR imaging data and machine learning (ML) to identify patients with vascular cognitive disease (VCD). VCD is currently underdiagnosed, and diagnosis is made at the late stages of the disease. At present, there are no treatments available for VCD, but if caught early, it is possible to stop the progression of the disease by, for example, lifestyle changes. This makes early detection of VCD highly important.&nbsp; 750 magnetic resonance imaging (MRI) images from a cohort of 506 women and men between the ages of 50 and 79 from the Gothenburg mild cognitive impairment (MCI) study were analysed. The cohort included controls, preclinical participants (subjective cognitive impairment and MCI), and patients with Alzheimer’s disease (AD), subcortical small-vessel disease (SSVD), and mixed AD/SSVD. FLAIR sequences from a 1.5 Tesla MRI scanner were used, together with already analyzed white matter hyperintensity (WMH) volumes acquired via the Freesurfer 5.3 software. The input to the model consisted of FLAIR MRI volumes, while the ground truth data included clinical diagnoses and WMH volume measurements. A multi-task learning (MTL) model based on ResNet18 was implemented, combining regression of WMH volumes and classification of patient diagnoses. The explainability method Guided Grad-CAM was implemented to visualize the regions of the MRI images that contributed most to the model’s predictions. The model was analyzed in a binary and two multiclass cases. The regression head produced WMH volume estimates comparable to those obtained from FreeSurfer. The classification head achieved an F1-score of 0.8414 for the binary case, but did not yield satisfactory results for the three-class and five-class cases, with F1-scores of 0.5381 and 0.4465, respectively. This thesis demonstrates that it is possible to quantify WMH volumes and identify patients with vascular cognitive impairment based on FLAIR images using ML.
Welcome!
Amanda Hoveklint &amp; Kajsa Homann

Title: Standards-Driven and AI-Driven Automation in Clinical Data Flows: A Case Study at AstraZeneca on Streamlining the Clinical Data Flow and Improving Interoperability within the Healthcare Industry
Presenter: Alexandra Johansson and Tove Thedin Olsson, MPMED Examiner: Stefan Candefjord Opponents: Anna Fassih and Elina Sandoff
Date and time: 5th of June 2025,15:00-16:00
Room: 3364 EDIT-rummet
Abstract: This study explores how standards-driven and AI-driven automation can streamline the clinical data flow at AstraZeneca (AZ) and improve interoperability within the healthcare industry. As clinical data volumes and complexity increase, efficient management of data is critical for accelerating drug development while ensuring robust clinical evidence. The study maps AZ’s current clinical data flow and identifies opportunities for automation and standardization. A qualitative research method was applied, using semi-structured interviews with AZ employees and internal documentation review. Process mapping and thematic analysis were conducted to generate a Data Flow Diagram, extract key insights, and deliver recommendations. Findings show that the current data flow involves several manual steps across data collection, management, statistical analysis, and regulatory submission. Key areas for improvement include stricter metadata requirements, a centralized data flow, and detailed workflow documentation to support automation. The study also highlights the potential of machine-readable specifications and executable metadata, aligned with the CDISC 360i program, to enable platform-based, streamlined data flows. Prioritizing critical-to-quality data in study protocols and shifting to structured, data-centric regulatory submissions are further recommended. Several AI opportunities were identified, including digital twins for virtual control arms, natural language processing for regulatory text generation, and agentic AI for end-to-end automation. These require strong quality control, human oversight, explainability, and regulatory acceptance. The study concludes that by proactively adopting these recommendations, AZ can improve internal efficiency, support a connected and data-driven healthcare ecosystem, and help lead the transformation toward faster, higher-quality drug development and timely patient access to innovative treatments. 
Welcome! Alexandra and Tove

Titel: Mechanisms to Mitigate Distribution Network Congestion from Solar Power Peaks
Student: Viktor Rodin Examiner: Ola Carlson
Date: 10 of June at 9:00 Room: Uno Lamm, 3571
Inclusion of rising rooftop photovoltaics in Swedish distribution networks introduces operational challenges such as voltage instability and constrained hosting capacity, necessitating targeted voltage mitigation measures. By integrating voltage-models for cables and transformers with PV sites site-specific data and PV site peak production forecasts, this thesis assess mitigation strategies that enhance hosting capacity and ensure voltage stability. The work is done for smaller network within the Falu Elnät distribution network.
Most Welcome! Viktor and Ola&nbsp;

Title: Machine learning-based identification of control algorithm for Chalmers test wind turbine
Student: Korin Damian Klein Examiner: Ola Carlson
Date: 10 of June at 10:00 Room: Uno Lamm, 3571
Abstract: Electricity generation from wind energy is one of the most promising renewable energy sources, with a rapidly growing share in the power grid. To ensure both optimal power output and safe operation, control systems play a critical role in wind turbines. Chalmers University of Technology conducts research using its own test turbine, located on the island of Björkö of the coast of Gothenburg, which is equipped with numerous sensors. Based on the analysis of measurement data from these sensors, this thesis develops a graphical method to identify the turbine’s control algorithm using a torque–rotor speed diagram. Methods for detecting changes in the control algorithm and estimating applied control parameters are also considered. The applicability of basic machine learning techniques to support these methods is evaluated.
Most Welcome Korin Damian Klein, Håkan Johansson and Ola Carlson

Title: Design and Simulation of a Microcontroller- Based Class D Power Amplifier for Piezoelectric Actuators in Bone Conduction Hearing Aids.&nbsp; Students: Albin Remnesjö Bergstrand and Emil Johansson. Examiner: Erik Agrell. Opponent: Jose Eduardo Maciel Mendoza. Place: E2 Room 7430 Landahlsrummet. Time: 10:00-11:00.
Abstract: Conventional linear audio power amplifiers have been and continue to be popular in audio applications. In recent years, digital class D switching power amplifiers have increased in popularity, as the demand for smaller and more efficient electronic devices is increasing. Together with advancements in technology, the quality of Class D amplifiers continues to improve. One hearing aid technology that uses Class D amplifiers is bone conduction hearing aids. These systems generally consist of an extrinsic microphone and sound processor, connected to a percutaneous device with the purpose of generating sound through vibrations to the skull bone. Historically, bone conduction devices have been designed and operated based on electromagnetic principles using electromagnetic actuators. Over time, a more efficient type of actuator is increasing in popularity, one that utilizes the piezoelectric force. Low-voltage Class D amplifiers are often only available in integrated circuits (ICs), which take a long time to develop and can be a bottleneck in the process. This thesis explores an alternative approach by developing a design and simulation framework for a low-voltage class D power amplifier using discrete components to drive a piezoelectric actuator load. The aim of the following approach is to reduce the development time while showcasing the feasibility of constructing low-voltage class D power amplifiers utilizing discrete components for piezoelectric actuator loads, achieving performance comparable to integrated solutions. A class D amplifier, using a H-bridge power stage, with ∆Σ to PWM with dead-time control, has been developed in simulation and with PCB design. The amplifier, operating across the audio frequency range (20 Hz to 20 kHz), achieved a peak SNR of 77 dB and a THD+N of 0.15% when driven with a 0.658 VRMS sinusoidal input with frequency 4410 Hz. A peak apparent efficiency of 86.4% was observed with an input voltage of 2.44 VRMS. Compared to IC designs reported in technical literature and found in commercial products, the results are promising. Thus, the conclusion is that the idea of designing a low-voltage discrete Class D amplifier for piezoelectric hearing aids is deemed viable. Still, its performance could be greatly improved by incorporating gate drivers, a higher-order ∆Σ modulator, and more refined filter designs.
Welcome!
Albin, Emil and Erik

Title: Automated material testing Student: Lorenzo Bosio Examiner: Martin Fabian Opponent: Giovanni Martini Time: Friday, 13 June, 13:15 Place: E2 Room 2503 Fredrik Lamm
Abstract: The integration of robotic systems into materials testing laboratories is increasingly recognized for its potential to enhance throughput and achieve high-precision experimental measurements. At the Volvo Cars facility, an automated 3-point bending test cell was developed to enable high-throughput material testing in accordance with the VDA 238-100 standard, ensuring consistent and accurate test execution. This project addresses key challenges encountered during system development, including automation integration, sample handling, and data management. Various design strategies and technical solutions are explored to optimize testing efficiency and measurement accuracy. Additionally, the potential for extending the system to accommodate other test types, such as tensile testing, is examined. The outcomes, challenges, and proposed enhancements provide valuable insights into advancing automated testing systems for industrial applications.
Welcome!
Lorenzo and Martin

Architecture and Civil Engineering

Bachelor’s thesis Chemistry and Chemical Engineering

Chemistry and Chemical Engineering

Communication and Learning in Science

Computer Science and Engineering

Industrial and Materials Science

Life Sciences

Mathematical Sciences

Mechanics and Maritime Sciences

Microtechnology and Nanoscience

Physics

Space, Earth and Environment

Technology Management and Economics

Presentation of master thesis

Choose subject

Course syllabus for

Subject

Each academic year contains different courses

See programme page at universityadmissions.se

Educational area

Permissible substitutions of compulsory courses

To see your programme syllabus – select the academic year in which you started the programme.

[en: page.studyprogrammepage.syllabusesview.coursesdescription]

Master thesis presentations at Electrical Engineering

26 May, 2025

10.00-11.00 Andreas Führ, MPCAS

14:00-15:00 Gideon Geelnard & Nandana Suryadevara, MPSYS

15.00-16.00 Anton Apelberg & Tom Ekman, MPEPO

15.00-16.00 Richard Nunstedt & Erik Linghammar, MPMED and MPENM

27 May, 2025

09.00-10.00 Diyasha Chenali Hapuarachchi, MPEPO

09:30-10:30 Axel Lundkvist, MPCAS

14.00-15.00 Fredrik Berntsson & Anna Oskarsson, MPSES

16.00-17.00 David Einarsson, MPEPO

28 May, 2025

10.00-11.00 Alva Cider & Norea Ronnby, MPMED

10.00-11.00 Fredrik Nygren & Oscar Schyum, MPPDE and MPSYS

11:00-12:00 Annie Abrahamsson Eknefelt & Ludvig Swala, MPMED

13.15-14.00 Nils Larsson & Daniel Svensson, MPSYS and MPSES

16.00-17.00 Pauline Nässlander & Anni Carlén, MPSYS

30 May, 2025

13.15-14.15 Johan Gustafsson & Dennis Hjertén, MPSYS

2 June, 2025

10.00-11.00 Elvira Carlén & Elsa Thoreström, MPMED

11.00-12.00 Anton Carlsson & Hugo Rådegård, MPCAS

13.00-14.00 Anish Janardhan & Muddur Ajay Nayak, MPSYS

14.00-15.00 Fatemeh Alibakhshi & Pawel Kasinski, MPENM and MPALG

16.00-17.00 Petter Dittmer, MPENM

3 June, 2025

09.00-10.00 Nowroz Bin Nasim & Kristoffer Olsson, MPEPO

13:00-14:00 Oskar Andersson & Ludvig Fornstedt, MPCAS

4 June, 2025

08.30-09.30 Nathalie Pettersson, MPSES

09.30-10.30 Amelia Bergum & Alida Nilsson, MPSES

09.00-10.00 Anthon Odengard & Anton Heden, MPCAS

09:00-10:00 Anna Fassih & Elina Sandoff, MPBME

10:30-11:30 Erik Blomqvist, MPMED

13.45-14.30 August Blanc, MPCAS

5 June, 2025

10:30-11:30 Amanda Hoveklint & Kajsa Homann, MPMED

15:00- 16:00 Alexandra Johansson & Tove Thedin Olsson, MPMED

10 June, 2025

09.00-10.00 Viktor Rodin, MPSES

10.00-11.00 Korin Damian Klein, MPEPO

11 June, 2025

10.00-11.00 Albin Remnesjö Bergstrand & Emil Johansson, MPBME

13 June, 2025

13.15-14.15 Lorenzo Bosio, Erasmus