Masterarbeiten und Dissertationen
A comparison of haptic and visual support for navigation in an audio-based city game
Tkacz, Jakub
Masterarbeit
Informaatioteknologian ja viestinnän tiedekunta – Faculty of Information Technology and Communication Sciences; Tampere University, 2024
Betreuer an der Universität Tampere: Dr. Ahmed Farooq und Prof. Mark van Gils
Betreuer am LBI-DHP: Jan Smeddinck, Devender Kumar
Abstract:
Navigation is a fundamental aspect of human life, yet current solutions are designed for vehicles, leaving pedestrian navigation as a derivative of its car equivalent. This thesis investigates a novel approach to pedestrian navigation, using haptic feedback to provide directional hints. Three methods were compared within navigation in an active city game: a haptic belt equipped with 16 vibrators, a smartwatch used as a haptic compass, and a smartphone with a compass displayed on the screen used as a baseline. A user study was used to evaluate these devices on four dimensions – comfort, performance, intuitiveness, and preference, gathering data via questionnaires, interviews, and logs from the devices. The results indicate that the visual compass, outperformed other methods in most dimensions, offering familiar experience and precise directions. Moreover, it was least prone to GPS inaccuracies by offering constant feedback. The haptic devices, even though showing lower results, were offering the highest environmental awareness, which may be beneficial in different scenarios or locations. Based on the experience from the user study, the multi-modal StageNav framework was introduced. It divides navigation into five distinct stages, recommending different feedback methods for each phase. This framework offers flexibility for designers and ensures that navigation offers the highest performance. The primary goal of the framework is to enhance the experience and safety of pedestrian navigation.
A Digital Talk Test for Assessing Exercise Intensity of Patients with Cardiovascular Diseases
Geiger, Laura
Masterarbeit
Universität Innsbruck (Department of Computer Science), 04.03.2024
Betreuer (UIBK): Ass.-Prof. Dipl.-Ing. Clemens Sauerwein, PhD
Co-Betreuer:innen (LBI DHP): Mag. Dr. Daniela Wurhofer, Bakk. techn., Dr. Devender Kumar, Dr.-Ing. Jan David Smeddinck, BSc, MSc
Abstract
Addressing cardiovascular diseases as a significant global health concern requires innovative approaches to enhance patient care and rehabilitation. One of these approaches is the development of a digital version of the Talk Test which enables patients to self-assess their intensity through speech during workouts. Furthermore, the automated assessment of exercise intensity through speech related to the Talk Test offers a compelling and engaging research area. Currently, no publicly available machine learning approaches address the automated exercise intensity estimation.
In this thesis a prototype of the Digital Talk Test, called aktivtalk, is created and used in a comparative study to collect self-assessed voice samples. The usability of the aktivtalk application achieves a higher user satisfaction compared to the non digital approach in the study. Furthermore, an initial machine learning model is created to predict the exercise intensity zone without the need for self-assessment. The performance of the model with the self-assessed labels even slightly outper-
formed a traditional age and pulse based labeling approach. In summary, this thesis emphasizes the potential of the Digital Talk Test as a valuable tool for assessing and managing exercise intensity in patients with cardiovascular diseases, highlighting its reliability, usability, and potential for future development in digital healthcare.
Social activities online – integrating social elements in a digital health app to support a heart-healthy lifestyle
Kolosovskaia, Daria
Masterarbeit
Paris Lodron Universität Salzburg und Fachhochschule Salzburg (Joint Master Programme on Human-Computer Interaction), 01.02.2024
Betreuerin: Dr. Daniela Wurhofer
Using AI-Generated Stories to Support a Heart-Healthy Lifestyle
Mayerhofer, Christoph
Masterarbeit
Fachhochschule Salzburg (Multi Media Technology), 25.01.2024
Betreuerin: Dr. Daniela Wurhofer
Towards a warning system for inaccuracies in PPG-based heart rate measurements on smartwatches : a supervised deep learning approach
Brunner, Marlene
Masterarbeit
Paris Lodron Universität Salzburg, Fakultät für Digitale und Analytische Wissenschaften, 14.09.2023
Betreuer:innen: Univ.-Prof. Dr. Arne Bathke, Dr.-Ing. Jan Smeddinck BSc, MSc, Dr. Devender Kumar
Abstract:
Wearable devices that measure Heart Rate (HR) using Photoplethysmography (PPG), such as smartwatches or fitness bands, have become increasingly popular in recent years. However, the accuracy of PPG-based HR measurements can be affected by a number of factors. This thesis investigates the accuracy and validity of PPG-based HR measurements in comparison to gold standard Elektrocardiogram (ECG) readings. First background literature is summarized to investigate the current state of the art in PPG-based HR measurement. Then, a real world data set is analyzed to assess the accuracy of PPG-based HR measurements in a real-world setting. The first part of the thesis shows that PPG-based HR measurements can be inaccurate, particularly at high exercise intensities. To address this issue, a Linear Regression model and two Deep Learning models have been developed to predict the measurement errors that occur based on the data stream from the wearable alone. The results show that a Deep Learning model based on a Convolutional Neural Network (CNN) outperforms the other models. The model is able to reliably detect large measurement errors and therefore it is possible to develop a warning system to inform the end user of the wearable when those large deviations from the gold standard occur.
A System for Executable Semantic Models of Just-In-Time Adaptive Interventions
Kremser, Wolfgang
Masterarbeit
Paris Lodron Universität Salzburg, Fakultät für Digital and Analytical Sciences, August 2023
Betreuerin: Assoz. Prof. Dipl.-Ing. Dr. Andreas Naderlinger
Abstract:
Mobile health, or mHealth, uses mobile apps to provide medical care, promote health, and manage disease. The goal is to cost-effectively monitor and improve the user’s physical condition by providing information and education, as well as tools for self-assessment and positive behavior change support. One example of an mHealth app is Aktivplan. This web-based mobile app allows healthcare professionals, together with their patients, to schedule physical activity designed to improve cardiovascular health. The patient can log the amount and intensity of physical activity per day using their personal smartphone. The app displays the patients’ activity levels to themselves and their healthcare professional for monitoring and regular checkups.
Intervention design is a central aspect of mHealth applications. It determines how an mHealth system reacts to changes in the user’s state. An emerging intervention design is the just-in-time adaptive intervention (JITAI). JITAIs are interventions of ’the right type and amount’ which are delivered ’at the right time’. They are an ongoing topic of research in the field of mHealth. To
better understand the effects of customization parameters, like intervention timing, mode and content, researchers conduct micro-randomized trials, in which variations of these parameters are tested.
This field of research is at the intersection of medical research and computer science. Study investigators are often medical experts and rely on application developers to implement JITAIs according to their study design. Changes in the study design thus necessarily cause friction, as these changes must be communicated and implemented.
This work aims to mitigate this friction by providing a comprehensive description of the different components required for an ontology-based system to realize ’codeless’ JITAIs. It presents several results that contribute to this goal: (1) the Aktivgraph ontology with which mHealth researchers can independently model JITAIs, (2) a software library that extends the existing Aktivplan app for the creation of Aktivgraph models, (3) the JITAI Engine to execute the semantics of Aktivgraph models, and (4) a demonstration of these results with a JITAI that sends patients push messages with timeslots where the weather is fit for activity. The demonstration shows that the JITAI Engine is capable of executing the semantics of an Aktivgraph JITAI model.
Utilising audio storytelling and gamification to promote outdoor physical activity
Veikoum, Dionysios
Masterarbeit
Paris Lodron Universität Salzburg und Fachhochschule Salzburg (Joint Master Programme on Human-Computer Interaction), 05.05.2023
Betreuerin: Dr. Daniela Wurhofer
Implementation and evaluation of an application that suggests physical exercises to bridge waiting times in everyday life
Maislinger, Magdalena
Masterarbeit
Fachhochschule Salzburg (Multi Media Technology), Puch bei Hallein, September 2022
Betreuerin: Dr. Daniela Wurhofer
Einsatz digitaler Technologien in der kardiologischen Rehabilitation aus der Sicht österreichischer Gesundheitsexpert:innen – Entwicklung einer Onlinebefragung
Lunz, Luisa
Masterarbeit
Paris Lodron University Salzburg (Sport- und Bewegungswissenschaft), Salzburg, April 2022
Betreuerin: Assoz. Prof. Dr. Sabine Würth
Ko-Betreuer: Dr. Tino Kulnik
Technology as bridge between health professionals and patients
Neunteufel, Julia
Masterarbeit
Paris Lodron Universität Salzburg and Fachhochschule Salzburg (Joint Master Programme on Human-Computer Interaction), Salzburg, 05.09.2021
Betreuer:innen: Dr. Daniela Wurhofer & Dr. Bernhard Maurer
Zusammenfassung:
Ziel: Der Zweck dieser Studie ist es, zu erforschen, ob und wie Technologie zur gemeinsamen Rehabilitationsplanung durch medizinisches Fachpersonal und PatientInnen beitragen kann.
Hintergrund: Weltweit stellen Herz-Kreislauf-Erkrankungen unverändert die häufigste Todesursache dar und sind somit eine Herausforderung für das Gesundheitswesen in Bezug auf Personal und Kosten. Rehabilitation und die Unterstützung der PatientInnen, langfristig aktiver zu werden, sind in diesem Zusammenhang wichtige Aspekte. Technologie findet hier verstärkt Anwendung. Die Gestaltung einer solchen Technologie ist noch in der Erforschung. Die Verwendung eines personenzentrierten Ansatzes in Verbindung mit Theorien aus dem digitalen Gesundheitsbereich kann für das Thema von Nutzen sein.
Herangehensweise: In dieser Studie nahmen fünf RehabilitationsexpertInnen und sieben PatientInnen teil. Dabei wurden semi-strukturierte Interviews, Beobachtungen und eine 14-tägige Tagebuchphase durchgeführt, um die Bedeutung von Technologie im Kontext der Rehabilitationsplanung zu untersuchen. Zur Auswertung der Daten wurde eine thematische Analyse verwendet.
Ergebnisse und Diskussion: Es konnten mehrere Faktoren, die die Einhaltung der vorgeschriebenen Aktivitäten beeinflussen, festgestellt werden. Vor allem die Motivation, die Ziele und der körperliche Zustand der PatientInnen konnten als solche Faktoren definiert werden. Die Technologie half bei der Erstellung maßgeschneiderter Trainingspläne und positive Auswirkungen auf das Gesundheitsverhalten der Teilnehmenden konnten festgestellt werden. Bei der Umsetzung des Trainings selbst wurde das digitale Tool als wenig unterstützend empfunden. In dieser Studie hat sich herausgestellt, dass Automatisierung, wie etwa das automatische Aufzeichnen der Aktivitäten durch andere Geräte, für die BenutzerInnen wünschenswert ist. Darüber hinaus schien es wichtig zu sein, sich realistische Ziele zu setzen, was durch eine passende Abbildung der Ziele in einer digitalen Anwendung unterstützt werden kann.
Fazit: Die Implementierung von mHealth im klinischen Umfeld birgt das Potenzial für eine effiziente und personenzentrierte Behandlungsplanung. Die Verwendung geeigneter Theorien aus Disziplinen wie der Psychologie sollte einer der treibenden Bestandteile in der Entwurfsphase eines solchen Produkts sein. Darüber hinaus ist aber auch ein geeignetes Konzept für die Implementierung der Technologie in die Organisation wichtig. Des Weiteren können das Setzen passender Ziele und das automatische Aufzeichnen der sportlichen Aktivitäten den Erfolg von mHealth Produkten steigern. Die Rückmeldung über die Aktivitäten der PatientInnen durch die Fachkräfte in der App, aber auch persönlich, ist wichtig für das Engagement der PatientInnen für ihre Behandlung.
Schlüsselwörter: medizinische Informatik, personen-zentrierte Medizin, gemeinsame Entscheidungsfindung, persuasive technology, Verhaltensänderung, mHealth, Ethnographie, HCI im Designen von mHealth