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

Design Features and Clinical Considerations of Orthopedic Knee Rehabilitation Exoskeletons: A Systematic Literature Review and Analysis
(2026) Wei, Zikun
As rehabilitation robotics expands from neurorehabilitation to orthopedics, the biomechanical suitability of existing devices remains unclear. This paper systematically reviews 57 wearable knee exoskeletons (2015–2025) to evaluate their clinical compatibility. It highlights the critical design divergence between the “repurposing strategy” of neuro-derived systems and the “specialized design philosophy” of orthopedic-specific solutions. Using a novel five-dimensional evaluation framework, this study reveals a significant design "mismatch." While systems relying on a generalist, neuro-adapted approach offer advanced control, their architectures often lack the kinematic compatibility and structural support required for vulnerable orthopedic joints. In contrast, "specialized" designs have seen explosive growth, prioritizing unilateral modularity and anatomical protection. However, a "valley of death" persists in translational research: over 60% of orthopedic-specific devices remain at Technology Readiness Level (TRL) 4, failing to reach clinical patient trials. The study concluded that most current cases of transplanting neurological rehabilitation equipment for orthopedic use cannot address the specific needs and limitations of orthopedics. Future development should adopt an "orthopedic-first" philosophy— prioritizing multi-DOF stability, modular adaptability, and pain-aware control. These shifts are imperative to reduce adverse events and optimize treatment outcomes in orthopedic rehabilitation. In addition, another systematic review should be conducted on mature and cuttingedge neurosurgical exoskeletons that do not claim to include orthopedic rehabilitation functions to analyze their potential to meet orthopedic rehabilitation needs.
Design and Development of an adaptive robotic rehabilitation exoskeleton for lower limb
(2026) Shen, Pengcheng
This thesis focuses on the design and development of an adaptive robotic rehabilitation exoskeleton for lower-limb assistance. The main objective of the project is to create a more compact, integrated, and portable electronic system that can be easily embedded into lower-limb exoskeleton structures for practical rehabilitation use. To achieve this goal, a custom electronic architecture based on the ESP32 Feather S3 and Arduino Portenta H7 microcontrollers was developed. The work includes the design of a custom expansion PCB, two adapter boards for MCU compatibility, and a fully customized IMU module for real-time motion sensing. These hardware components enable improved functionality, reduced system size, and enhanced structural integration. In addition, dedicated firmware was implemented to ensure stable communication, control, and data acquisition across the exoskeleton system.
Antibody-mediated targeting of NOX2 as a potential therapeutic strategy in cancer
(2026) Ankarberg, Elin
NADPH oxidase 2 (NOX2) is a key enzyme in phagocytic microbial killing through its production of superoxide (O2•−), which can subsequently be converted into other reactive oxygen species (ROS). In monocytes and other myeloid cells, NOX2 is localized to endolysosomal and plasma membranes, enabling the generation of both intra- and extracellular ROS. Although ROS are protective in some contexts, excessive extracellular ROS may suppress cancer-targeting lymphocytes, including natural killer (NK) cells, thereby contributing to cancer initiation and progression. In mouse models, both pharmacological and genetic inhibition of NOX2 reduces tumour burden and metastasis, and NOX2 inhibitors have been explored clinically in leukaemia and other cancers. Monoclonal antibodies are powerful therapeutic agents and may represent a promising, yet understudied, strategy for targeting NOX2. This thesis aimed to evaluate the effects of an anti-NOX2 antibody, 7D5, on extracellular ROS production in human monocytes using isoluminol-enhanced chemiluminescence assays. Both 7D5 and a control antibody tended to reduce extracellular ROS, although 7D5 was slightly more effective when ROS production was stimulated by a bacterial peptide. We also examined whether 7D5 could uphold NK cell viability and function in co-cultures of NK cells, monocytes, and the K562 leukemic cell line. Unlike the control antibody, 7D5 significantly protected NK cells from ROS-induced apoptosis. In summary, a monoclonal antibody against NOX2 weakly affected ROS production from human monocytes and yet preserved NK cell viability in co-cultures with ROS-producing monocytes. NOX2 antibodies may thus represent a strategy to improve NK cell function in cancer immunotherapy.
On-Chip Bandpass Filter for Superconducting Devices
(2026) Winkel, Job
On-chip lumped element bandpass filters offer a pathway to tightly integrate noise suppression directly at the chip level in superconducting quantum devices. Despite the widespread use of filters in cryogenic qubit setups, co-fabricated lumped element bandpass filters remain relatively unexplored. This work evaluates their feasibility, design constraints, and performance when embedded directly on a superconducting qubit chip, paving the way for scalable quantum architectures. The filter design follows a standard radio frequency (RF) synthesis approach, adapted to cryogenic operation, co-fabrication constraints, limited footprint, and superconducting drive requirements. A scalable design flow is developed to implement arbitrary-order bandpass filters using lumped inductors and capacitors. Electromagnetic (EM) simulations are employed to extract effective component parameters and refine circuit models beyond ideal lumped element approximations. Simulations show that on-chip lumped element bandpass filters can achieve welldefined passband characteristics and support higher-order architectures. However, ideal and extended lumped element models alone are insufficient to predict device response accurately. Direct optimization with computationally intensive EM simulations were therefore necessary to achieve reliable filter performance. The filter response also directly influences the thermal noise spectrum experienced by the qubit. Modeling indicates that appropriately designed bandpass filters can reduce unwanted thermal occupation, providing a tool for engineering and investigating the qubit’s EM environment. A prototype device was fabricated and characterized at cryogenic temperatures. The measured response did not exhibit the intended passband, with analysis pointing to fabrication issues, particularly unreliable capacitor connections, rather than limitations of the filter concept or design methodology. Overall, this work establishes a simulation-driven platform for co-fabricated lumped element bandpass filters in superconducting quantum circuits. The results demonstrate their feasibility, scalability, and potential for controlled thermal noise engineering in cryogenic quantum hardware.
Retrieval-Augmented Generation for Sustainable Material Data Handling in Automotive Value Chain
(2026) Tran, John; Larsson, Daniel
Applying large language models (LLMs) to industrial material data workflows has the potential to improve efficiency. However, conventional LLMs are limited by hallucinations, depend on proprietary training data, and are costly to update. This thesis explores Retrieval-Augmented Generation (RAG) as an alternative approach, in which an LLM generates responses grounded in a restricted, domain-specific corpus of documents and databases and provides source citations for them. The study is carried out in an industrial setting with two main data domains: an internal SQL materials database with tabular material properties, and a corpus of unstructured textual documents, including supplier documents, corporate standards, and environmental product declarations. A RAG system is developed that (1) indexes both textual and tabular data, (2) retrieves relevant chunks via dense vector search, and (3) generates source-grounded responses. The work investigates whether a RAG model that explicitly integrates both domains can outperform a baseline tuned for unstructured text and explores which tabular serialization format yields the most semantically informative embeddings for pretrained embedding models. To achieve this, we first constructed LLM-based pipelines to generate documentand table-based test sets with ground-truth chunk annotations, and implemented a modular RAG pipeline with separate indices for textual and tabular data. Then, we experimented with multiple retrieval strategies, ranging from concatenating the retrieval results to using cross-encoders to weigh them. In addition, several fusion strategies were tested to evaluate whether they could improve retrieval accuracy when operating across different domains. Experiments are conducted comparing nine tabular serialization strategies, studying performance as a function of index size, chunk size, and top-k, and evaluating different fusion modes and embedding models. The evaluation metrics used are Hit Rate, Recall, Precision, F1-score, and Mean Reciprocal Rank. Results show that enriched serialization, which converts tabular rows into natural-language statements, yields stronger tabular retrieval performance than a standard key-value-based format, without degrading performance on document retrieval. Larger chunk sizes and higher top-k values systematically improve retrieval metrics, highlighting both the difficulty of relying solely on similarity search and the benefits of cross-encoder reranking on larger candidate sets. A domain-aware weighted fusion retriever further improves overall retrieval performance over the optimized baseline with only moderate computational overhead. These findings demonstrate that semantically rich tabular representations and domain-aware fusion can enhance RAG performance on heterogeneous industrial material data.