Examensarbeten för masterexamen
Länka till denna samling:
Browse
Senast publicerade
Visar 1 - 5 av 1755
- PostRecommendations for Presenting Battery State of Health Information - To enhance in-vehicle experience and understanding for Battery Electric Vehicle Users(2024) PONGCHANCHAI, NATTHAMON; TUMMA, LIINA; Chalmers tekniska högskola / Institutionen för data och informationsteknik; Chalmers University of Technology / Department of Computer Science and Engineering; Fjeld, Morten; Obaid, MohammadWithin the automotive industry nowadays, a significant transformation is currently happening, transitioning from combustion engine vehicles to more sustainable alternatives, including Plug-in Hybrid Electric Vehicles (PHEVs) and Battery Electric Vehicles (BEVs). With more and more consumers coming to the fully-electric vehicle market, the new regulations concerning the battery State of Health (SoH) of Electric Vehicles (EVs) have been regulated and used as a standardisation for automotive manufacturers to implement battery SoH feature by 2026. This new legislation is not only new to the EV drivers but also novel to the vehicle manufacturers in the automotive industry. However, compliance with the regulations does not take into account the user experience. Moreover, the scarcity of research in this area makes it potentially challenging to understand the user needs and goals. This thesis investigates how the information of battery SoH should be displayed from a user-centred perspective, with the aim to improve user understanding of battery health and increase their engagement with the BEV battery health maintenance. The study takes a user-centred design approach to better understand user needs and develop the most appropriate approach to display battery SoH to drivers. During the process, several research and design methodologies including literature review, questionnaire and focus groups, have been conducted to understand the users, frame the problem spaces, and identify the user requirements for further development. With the defined user requirement list, different resolution levels of prototypes were developed and evaluated in three iterations, with a total of 15 participants involved across these iterations, to ensure that the design concept met their expectations and requirements and that the prototypes were refined efficiently based on their feedback. The findings revealed that simply displaying SoH as a percentage is insufficient to inform users about the current state of their battery health and additional contextual information and actionable guidance are needed. This thesis presents a series of design recommendations as deliverables, derived from research, analysis, and evaluation of the design concept. These recommendations were established with the aim of enhancing the user experience and understanding of battery SoH information displayed in BEVs, also serving as guidance for the implementation of the SoH functionality in the BEVs in the upcoming future.
- PostAbandoned Object Detection in Vehicle Cabin Environment(2024) Hu, Hao; Cheng, Yuhua; Chalmers tekniska högskola / Institutionen för data och informationsteknik; Chalmers University of Technology / Department of Computer Science and Engineering; Petersen Moura Trancoso, Pedro; Petersen Moura Trancoso, PedroAbandoned object detection in the vehicle cabin environment is pivotal for the convenience of passengers. Existing methods for abandoned object detection are either aimed at outdoor surveillance or rely too much on training data. This study aims to develop an efficient algorithm for detecting abandoned objects in a vehicle cabin environment that relies less on training data and generalizes to untrained objects. The main approach involves utilizing a simple background model to extract candidate abandoned objects, followed by feature extraction using MobileNetV2 pre-trained on ImageNet-k for accurate detection. By comparing features before and after human presence, false positive proposals are effectively filtered out. To make the proposed method suitable for embedded execution, optimizations are performed to improve efficiency. We evaluate the proposed method on common left-behind objects videos with a Jetson Nano device. The results prove the efficacy and efficiency of the proposed method. However, the study is still limited by the lack of research on dynamic cameras, low-contrast infrared images, and persistent shadows.
- PostStereoscopic Depth Imaging for Improved Point Cloud Generation of Skeletal Surfaces in IR-guided Orthopaedic Surgery(2024) Cedervall, Mats; Dahlqvist, Oscar; Chalmers tekniska högskola / Institutionen för data och informationsteknik; Chalmers University of Technology / Department of Computer Science and Engineering; Sintorn, Erik; Assarsson, UlfThis thesis proposes a flexible approach of combining traditional optical navigation systems with a visible light-based depth camera for aligning preoperative CT models with the corresponding bone exposed during active surgery, providing real-time feedback of implant positioning to the surgeon based on a preoperative plan. By leveraging the Intel® RealSenseTM D405 depth camera, the study investigates the benefits and challenges of using stereoscopic depth imaging for 3D reconstruction of musculoskeletal surfaces. The primary goal being to evaluate how this technology can be used to reduce surgery duration in orthopaedic total knee arthroplasty (TKA) surgery. The proposed method transforms the depth images captured using the depth camera into a navigation system’s frame of reference, accomplished through a series of preoperative calibration steps. The reconstructed intraoperative 3D model is then aligned with the CT model using iterative closest point (ICP) algorithms. This thesis also includes an investigation into using light polarisation filters, analysis of hyperparameter tuning, and accuracy evaluation of the Intel® RealSenseTM D405 camera’s depth estimation capabilities. Experimental validation includes a mock surgery on pig cadaver parts to simulate intraoperative conditions. Results demonstrate that the proposed approach achieves good alignment accuracy at around 0.5 mm, though it exhibits higher variability compared to competing methods. In conclusion, the integration of this class of depth cameras with optical navigation systems is a viable solution for improving the speed of knee registration in orthopaedic surgery, but future work is encouraged to address depth estimation inaccuracies.
- PostDiscovering Error Handling Strategies for Improving End User Experience in Collaborative Distributed Real Time Systems(2024) Andreasson, David; Gunnarsson, Karl; Chalmers tekniska högskola / Institutionen för data och informationsteknik; Chalmers University of Technology / Department of Computer Science and Engineering; Fjeld, Morten; Torgersson, OlofThis thesis investigates error-handling strategies in collaborative distributed realtime systems, focusing on enhancing the end-user experience. Collaborative environments like Yolean, characterized by complex distributed architectures, often suffer from errors that negatively impact user satisfaction and system functionality. Through a combination of Research Through Design, Requirements Engineering, and Design Thinking, this study identifies error patterns and evaluates user preferences for managing these errors. Our research involved semi-structured interviews, affinity diagramming, storyboarding, prototyping, and iterative user testing to develop an understanding of user expectations and effective error-handling strategies. We developed and refined prototypes addressing two main scenarios: Distributed architecture and collaborative interactions with human-induced errors. The proposed error-handling strategies aim to provide intuitive and non-disruptive user support, enhancing system usability and reliability. The findings from the study are summarized as an initial framework to guide the exploration of error-handling strategies in such environments. Results from the study indicate that the design artifact improved the end-user experience compared to the current state of the Yolean application. However, further research would be needed to draw generalized conclusions and guide further refinement of the proposed framework.
- PostIn-car voice assistants: The need and potential for AI-enabled voice assistants in vehicles(2024) LIND JONSSON, AMANDA; HÖGMAN MÖLLER, ELIN; Chalmers tekniska högskola / Institutionen för data och informationsteknik; Chalmers University of Technology / Department of Computer Science and Engineering; Fjeld, Morten; Torre, IlariaThis thesis investigates the role of AI-enabled voice assistants within the automotive industry, with a focus on user needs and regional differences between the United States and Sweden. Utilizing a mixed-method approach inspired by the TripTech method, this study collected both quantitative and qualitative data through surveys and expert interviews. The results revealed that more than half of the participants viewed in-car voice assistants (VAs) positively, valuing their safety and convenience for handling tasks and accessing information while driving. However, qualitative insights highlighted flaws in language comprehension and accuracy in current automated dialogue system, such as misdialed contacts due to misunderstood commands. While users from the United States generally perceived less importance in problem scenarios and were less likely to use VAs compared to Swedish users, the potential for generative AI to enhance VAs was recognized. This could lead to improved comprehension and more dynamic user interactions, although concerns about privacy and response accuracy persist. The findings suggest that while generative AI offers promising enhancements for in-car VAs, significant challenges in privacy and accuracy need to be addressed to fully leverage this technology in vehicles.