Open Theses

XRanatomy: Foundations of a Modular Multisensory XR Atlas of the Human Body

This thesis lays the groundwork for a modular, multisensory XR atlas of the human body. The long-term vision is an extensible platform in which different body regions—such as the ear, the heart, blood vessels [2], or other organs—can be explored through rich combinations of visual, auditory, and additional sensory cues. Beyond healthy anatomy, the platform is intended to accommodate later modules that simulate medical conditions, functional impairments (e.g., visual [1] or auditory restrictions), and other clinically relevant variations. Rather than implementing a single, monolithic experience, this work focuses on designing and realizing the core architecture and interfaces that allow independently developed “experience modules” to plug into a shared XR framework.

The project will (1) define requirements for a multisensorial XR anatomy platform in collaboration with domain stakeholders, (2) design a software architecture that separates a central host application from self-contained body-part modules, and (3) implement clear interfaces for data exchange, interaction logic, and multimodal feedback (e.g., 3D anatomy visualization, spatial audio, and optional haptic or other sensory channels). As a proof of concept, the thesis will realize at least one complete module—for example, an organ that can be explored with coordinated visual and acoustic feedback—and integrate it into the host system.

The expected contribution is a reusable XR infrastructure, including documentation and example implementations, that can serve as a foundation for future work on a multisensory XR atlas of the human body. This enables multiple researchers and designers to develop diverse anatomical experiences in parallel, while maintaining technical compatibility, extensibility, and a coherent user experience.

[1] K. Krösl, C. Elvezio, M. Hürbe, S. Karst, S. Feiner and M. Wimmer, "XREye: Simulating Visual Impairments in Eye-Tracked XR," 2020 IEEE Conference on Virtual Reality and 3D User Interfaces Abstracts and Workshops (VRW), Atlanta, GA, USA, 2020, pp. 830-831, doi: 10.1109/VRW50115.2020.00266.

[2] B. Hentschel et al., "Interactive Blood Damage Analysis for Ventricular Assist Devices," in IEEE Transactions on Visualization and Computer Graphics, vol. 14, no. 6, pp. 1515-1522, Nov.-Dec. 2008, doi: 10.1109/TVCG.2008.142.

In cooperation with Stefania Serafin (DTU)

Exploring the Role of Acoustic Stimuli in Virtual Nature Experiences

While the restorative role of nature is relatively well understood [1], little is known regarding the contribution of acoustic cues. This thesis investigates how carefully designed visual and acoustic stimuli shape restorative, nature-based experiences in Virtual Reality [2,3], with a particular focus on the multisensory mechanisms that underlie immersion, presence, and well-being. Grounded in research in spatial audio, ecoacoustics, and perceptual restoration, the project examines how sound interacts with visual fidelity to construct convincing virtual nature environments.

The goal is to explore whether—and under which conditions—sound can compensate for limited visual detail, amplify the experiential qualities of high-fidelity visuals, or even function as the dominant modality in evoking restorative responses. 

Using spatialized soundscapes, procedurally generated environmental audio, and controlled manipulations of audiovisual congruency, the work examines environments such as forests, lakesides, and mountain settings, or other environments where nature plays an important role.

The methodological approach combines:

• Spatialized sound design 

• Physically informed or procedurally sound synthesis of natural phenomena,

• Subjective and physiological measures of immersion, presence, and affective restoration.

Through these multisensory manipulations, the project aims to identify the perceptual cues that most strongly support restorative outcomes, particularly in situations where visual realism is constrained (e.g., lightweight mobile VR, therapeutic applications, or accessibility-driven designs). The findings contribute to a deeper understanding of how auditory information structures virtual nature experiences and offer actionable design guidelines for creating effective, resource-efficient, and inclusive VR applications for wellbeing, rehabilitation, and stress reduction.

[1] Kaplan, Stephen. "The restorative benefits of nature: Toward an integrative framework." Journal of Environmental Psychology 15.3 (1995): 169-182.

[2] Rachel Masters, Victoria Interrante, Madeline Watts, and Francisco Ortega. 2022. Virtual Nature: Investigating The Effect of Biomass on Immersive Virtual Reality Forest Bathing Applications For Stress Reduction. In ACM Symposium on Applied Perception 2022 (SAP '22). Association for Computing Machinery, New York, NY, USA, Article 6, 1–10. https://doi.org/10.1145/3548814.3551459

[3] Mostajeran, Fariba, et al. "Effects of exposure to immersive computer-generated virtual nature and control environments on affect and cognition." Scientific Reports 13.1 (2023): 220.

In cooperation with Stefania Serafin (DTU)

Enhancing Teleportation-Based Locomotion in Virtual Reality: Design, Implementation, and Evaluation of Journey-Aware Movement Techniques

This thesis investigates how teleportation—the predominant locomotion technique in Virtual Reality—can be augmented to better preserve users’ sense of distance, progression, and achievement during virtual journeys. While teleportation effectively minimizes cybersickness by removing continuous motion, it also eliminates experiential qualities central to travel, such as spatial awareness, path comprehension, and the feeling of “having gone somewhere.” The goal of this work is to explore, design, and empirically evaluate alternative or complementary mechanisms that enrich teleportation-based locomotion without compromising comfort. Possible application areas include gaming, but also professional immersive-technology contexts such as relaxation experiences, therapeutic VR, and virtual journeys where traversal itself is meaningful.

Teleportation is widely used across VR applications because it reduces motion-induced discomfort. However, its discontinuous nature undermines many aspects of situated movement, which are crucial in exploratory environments, training simulations, narrative VR, or any experience where travel itself carries meaning. This project contributes to ongoing research in immersive technologies, spatial perception, and VR interaction design by developing and testing novel locomotion concepts that reintroduce journey-related cues into otherwise instantaneous travel.

Objectives

• Analyse the limitations of standard teleportation techniques regarding spatial understanding, distance perception, and user engagement.

• Design one or more enhanced teleportation concepts (e.g., transitional micro-animations, path-based visualizations, auditory or haptic cues, pseudo-movement overlays, progressive “stepping” teleports).

• Implement the developed techniques in a Unity-based VR prototype with configurable parameters.

• Compare the enhanced locomotion variants against standard teleportation in a controlled user study, focusing on comfort, cognitive load, spatial awareness, and perceived sense of journey or accomplishment.

• Evaluate qualitative and quantitative outcomes and derive design implications for future locomotion systems.

Designing a Multi-User XR Commute Experience for Remote Work Research

Currently being carried out by Shuokai and Mathias Steen Jelshøj

This thesis focuses on the design, implementation, and evaluation of a virtual commute experience intended to support research on remote work and home-office wellbeing. The goal is to simulate a metaphorical daily commute—such as a train or bus ride—in Virtual Reality, where users encounter pseudo-random social interactions with other participants. The system will serve as an experimental platform for exploring how lightweight, structured transitional experiences might influence social connectedness, mood, and daily routines in remote-work contexts.

Remote work reduces spontaneous social encounters and informal transitions between home and workplace. Virtual commute concepts provide an opportunity to reintroduce these transitions in a controlled and customizable way. By enabling short, repeated VR experiences in everyday environments, the work will contribute to ongoing research at DTU on immersive technologies, human-computer interaction, and wellbeing.

Objectives

• Design a VR experience simulating a daily commute with varied environmental states and pseudo-random social encounters.

• Implement a multi-user infrastructure, enabling asynchronous or synchronous short interactions between participants across several days.

• Deploy the system in a field setting, allowing participants to experience the commute from their homes.

• Conduct an exploratory user study, capturing qualitative and quantitative insights into user perceptions, routines, and social connectedness.

• Analyse and report findings, contributing to broader discussions on VR for remote-work support.

A Little Bit Less Impossible Spaces: Perceptual Acceptance of Height-Encoded Impossible Spaces in Room-Scale Virtual Reality

This thesis investigates how the design of “impossible spaces” in room-scale virtual reality can be made more acceptable and less noticeable to users by leveraging virtual height differences. Impossible spaces [1] aim to overcome limited physical tracking areas by cleverly overlapping and reusing physical space: through the placement of corridors, doors, and turns, users can physically walk within a small room while experiencing a much larger virtual environment with locally inconsistent geometry. While prior work has focused mainly on horizontal redirection and layout tricks, the potential of the vertical dimension in explaining or masking geometric inconsistencies remains largely unexplored.

The project will implement a set of impossible-space layouts in a VR prototype that systematically vary how impossible overlaps are “explained” using height changes (e.g., stairs, ramps, multi-level rooms) versus traditional flat-floor designs. In a controlled user study, participants will explore these environments using natural walking. The study will compare conditions in terms of perceived spatial coherence, acceptance and trust in the environment, detection of inconsistencies, presence, and potential side effects such as discomfort or disorientation.

The goal of the thesis is to derive design guidelines for integrating vertical structures into impossible spaces, and to better understand when and how height can be used to make physically impossible virtual environments feel plausible and comfortable to users.

[1] E. A. Suma, Z. Lipps, S. Finkelstein, D. M. Krum and M. Bolas, "Impossible Spaces: Maximizing Natural Walking in Virtual Environments with Self-Overlapping Architecture," in IEEE Transactions on Visualization and Computer Graphics, vol. 18, no. 4, pp. 555-564, April 2012, doi: 10.1109/TVCG.2012.47

Beyond Egocentric View: The Role of Motor Synchrony and Visual Perspective in an Immersive Rowing Ergometer Setup

This thesis investigates whether an egocentric visual perspective [1] is indeed the dominant factor for presence in immersive applications, or whether other components—particularly motor synchrony, or interaction fidelity [2]—can play an equally or more important role. We study this question in a controlled rowing ergometer setup, in which a seated user rows in front of a large display that presents an immersive virtual environment. Within this scenario, we systematically manipulate visual perspective (e.g., first-person vs. third-person or offset viewpoints) and degrees of motor synchrony between the user’s physical rowing motion and the virtual movement of a boat or avatar.

The project will implement multiple experimental conditions that vary these parameters, while keeping other aspects of the experience constant. In a user study, participants will row under these conditions and report on their sense of presence, embodiment, spatial involvement, motivation, and engagement. Objective measures such as session duration, rowing intensity, or voluntary continuation could complement self-reports to better understand how visual perspective and motor synchrony jointly shape the experience.

The goal of the thesis is to disentangle the relative contributions of egocentric view and motor synchrony to presence and motivation in this type of setup, and to challenge the assumption that egocentric visuals are always optimal for immersive interaction. The resulting findings can inform the design of sports, rehabilitation, and exergaming applications that leverage non-egocentric perspectives without sacrificing presence or engagement.

[1] Slater M, Spanlang B, Sanchez-Vives MV, Blanke O (2010) First Person Experience of Body Transfer in Virtual Reality. PLOS ONE 5(5): e10564. https://doi.org/10.1371/journal.pone.0010564
[2] Martin Hedlund, Cristian Bogdan, Gerrit Meixner, and Andrii Matviienko. 2024. Rowing Beyond: A Demonstration of Steering Methods for Rowing-based Locomotion in Virtual Environments. In Extended Abstracts of the CHI Conference on Human Factors in Computing Systems (CHI EA '24). Association for Computing Machinery, New York, NY, USA, Article 413, 1–6. https://doi.org/10.1145/3613905.3648643

Emotion-Driven and Environment-Aware Augmented Reality Lava Lamp Using Biosensing and Machine Learning

This thesis explores the design and development of an emotion-driven, environment-aware augmented reality lava lamp that adaptively visualizes user state and spatial context. The work focuses on creating a processing pipeline that integrates biosignals—such as respiratory patterns—with extended world knowledge obtained from an AR headset, including 3D scene information. These data streams are fused and mapped into a dynamic 3D density representation using a machine-learning approach, for example, a self-organizing map [1]. The resulting volumetric field is visualized through techniques such as marching cubes to generate a fluid, continuously evolving AR lava-lamp experience. The project aims to demonstrate how physiological input and environmental factors can be combined to create expressive, affect-responsive visualizations in immersive interfaces.

[1] Stephanie Andrews, Jordan Beth Vincent, and John McCormick. 2015. Duet: Improvising Spatial Dialogues with an Artificially Intelligent Agent. In Proceedings of the 3rd ACM Symposium on Spatial User Interaction (SUI '15). Association for Computing Machinery, New York, NY, USA, 57–60. https://doi.org/10.1145/2788940.2788952

Leg Muscle Contraction as an Input Modality for Virtual Locomotion Interfaces

Convincing virtual locomotion for a broad range of users remains one of the central open challenges in virtual reality. Despite decades of research, existing locomotion interfaces still struggle to balance natural movement, usability, and negative side effects such as cybersickness or fatigue. This thesis explores leg muscle contraction as an input modality for virtual locomotion, with a focus on steering and walking-in-place interfaces [1] that may offer more convincing and accessible solutions than current approaches.

The project will investigate how signals related to muscle activity in the legs—captured, for example, via EMG sensors or alternative sensing technologies—can be mapped to locomotion control in VR. Based on this sensing pipeline, the work will implement and compare several locomotion techniques, such as muscle-based steering (e.g., leaning or asymmetric contraction for turning) and muscle-driven walking-in-place (e.g., rhythmic contraction patterns to trigger forward movement). In a user study, these techniques will be evaluated in terms of perceived naturalness, control accuracy, mental and physical effort, presence, and potential side effects, and compared against baseline methods such as controller-based or standard walking-in-place interactions.

The goal of the thesis is to assess the feasibility and design space of leg-muscle-based locomotion interfaces and to identify conditions under which this modality can provide more convincing, comfortable, and inclusive virtual locomotion than existing solutions.

[1] D. Zielasko, S. Horn, S. Freitag, B. Weyers and T. W. Kuhlen, "Evaluation of hands-free HMD-based navigation techniques for immersive data analysis," 2016 IEEE Symposium on 3D User Interfaces (3DUI), Greenville, SC, USA, 2016, pp. 113-119, doi: 10.1109/3DUI.2016.7460040.

In cooperation with Deren Y. Barsakcioglu (DTU)

Towards a Standardized Fitts-Law-Inspired Testbed for Evaluating Virtual Locomotion Interfaces

While Fitts’ Law [1] and its ISO-standardized variants have become the de facto benchmark for evaluating pointing and selection techniques, there is currently no comparable standard for assessing virtual locomotion interfaces. Existing approaches either adopt Fitts-style tasks almost directly—thus sacrificing ecological validity for users’ natural movement—or rely on highly bespoke navigation scenarios that are difficult to compare across studies. This thesis aims to develop and validate a locomotion testbed that strikes a balance between standardization and realism: it is conceptually grounded in the original Fitts’ Law selection task, yet incorporates more natural paths toward the next target location.

The project will design a set of repeatable navigation tasks in VR, in which participants move between targets along controlled but plausible paths using different locomotion techniques. Task parameters analogous to Fitts’ Law—such as path length and effective target width—will be systematically varied to define an “index of difficulty” for locomotion. A user study will evaluate multiple locomotion methods (e.g., joystick-based movement, teleportation, ...) in terms of speed, accuracy, path efficiency, reliability, and subjective measures such as workload, comfort, and presence.

The goal of the thesis is to derive a task design and set of metrics that are (a) standardized and repeatable across studies, (b) sensitive enough to reveal performance differences between locomotion techniques, and (c) sufficiently natural to maintain external validity. In doing so, it takes a step toward an ISO-style benchmark for locomotion interfaces that parallels the role Fitts’ Law plays for pointing and selection.

[1] Soukoreff, R. William, and I. Scott MacKenzie. "Towards a standard for pointing device evaluation, perspectives on 27 years of Fitts’ law research in HCI." International journal of human-computer studies 61.6 (2004): 751-789. https://doi.org/10.1016/j.ijhcs.2004.09.001

Subtle EMG-Based Confirmation for Gaze Interaction in Head-Mounted Displays

Gaze input is an emerging interaction modality for head-mounted displays (HMDs) that enables users to quickly select targets simply by looking at them. However, confirming these selections remains challenging when relying on gaze alone. Common alternatives such as handheld controllers or camera-based hand gestures (e.g., pinching [1, 2]) can be fast and intuitive but are not always accessible (e.g., for users with limited motor abilities), feasible (e.g., under occlusions, in low-light conditions, or in constrained spaces), or socially acceptable (e.g., in public environments). This thesis investigates how electromyography (EMG) can be used to provide robust, comfortable, and socially acceptable confirmation gestures for gaze-based interaction. Building on initial work that compared different EMG electrode placements and gesture types (such as finger tapping, foot tapping, and biting) and showed that EMG-based confirmations can perform comparably to camera-based pinching [3], the project will focus on making such techniques more practical and reliable for everyday use. Possible application areas include accessibility-oriented VR, professional and industrial HMD workflows, public or shared spaces (e.g., museums, exhibitions), and other scenarios where reliable, low-visibility confirmation of gaze selections is important. The thesis contributes to ongoing research in immersive technologies, accessibility, and multimodal interaction by systematically designing and evaluating EMG-based confirmation techniques for gaze input, with a focus on robustness over longer sessions, usability in different environments, and perceived comfort and acceptability

[1] Ken Pfeuffer, Benedikt Mayer, Diako Mardanbegi, and Hans Gellersen. 2017. Gaze + pinch interaction in virtual reality. In Proceedings of the 5th Symposium on Spatial User Interaction (SUI '17). Association for Computing Machinery, New York, NY, USA, 99–108. https://doi.org/10.1145/3131277.3132180
[2] K. Pfeuffer, H. Gellersen and M. Gonzalez-Franco, "Design Principles and Challenges for Gaze + Pinch Interaction in XR" in IEEE Computer Graphics and Applications, vol. 44, no. 03, pp. 74-81, May-June 2024, doi: 10.1109/MCG.2024.3382961.
[3] Judith Hartfill, Michael Arz, and Frank Steinicke. 2025. EMG-based Confirmation of Gaze Selection in Extended Reality. In Proceedings of the Mensch und Computer 2025 (MuC '25). Association for Computing Machinery, New York, NY, USA, 375–384. https://doi.org/10.1145/3743049.3743062

In cooperation with Deren Y. Barsakcioglu (DTU)

AI-Personalized Generative VR Worlds for Stress Reduction

This thesis investigates how AI-generated and AI-personalized virtual environments can support stress reduction and relaxation in Virtual Reality. VR is a strong medium for calming experiences (e.g., meditation and restorative environments) because it allows high experimental control and consistent exposure. At the same time, people differ substantially in what they perceive as calming. Building on recent advances in generative AI and modern computer graphics (e.g., neural rendering / neural scene representations), the thesis focuses on whether personalization improves restorative outcomes. Concretely, the project compares non-personalized generative VR worlds (e.g., generic prompts or curated baselines) against personalized worlds created from a participant’s own input (e.g., text prompts or guided preferences). As a proof of concept, the student will implement a VR prototype based on an existing generation pipeline in the lab and evaluate it in a controlled user study.

The study is planned to extend beyond subjective measures by assessing autonomic arousal and the sympathetic/parasympathetic balance (e.g., via heart-rate variability derived from a simple ECG).

Research question: Does AI-personalized generative VR lead to stronger relaxation/restoration than non-personalized generative VR?

Augmented Reality for Interactive Learning of Probability and Set Operations

Currently being carried out by Bibek Adhikari

Many students struggle with statistics. Abstract notions such as unions, intersections, complements, conditional probability, and event spaces often remain purely symbolic for them, making it difficult to interpret problems, understand what expressions like P(A|B) actually mean, and see how these concepts connect to real situations and data.

This thesis explores how immersive technologies — in particular, Augmented Reality (AR) — can support the learning of fundamental concepts in statistics. AR offers the possibility to embed such abstract structures directly into the physical environment, enabling interactive, spatial, and situated learning experiences.

The project investigates how probabilistic reasoning and set relationships can be represented and manipulated in an interactive AR system. Learners will be able to explore events, sample spaces, and set operations through spatial overlays anchored in the real world, allowing them to experiment with configurations, observe consequences of operations, and develop an intuitive understanding of abstract relationships. The thesis builds on principles of situated learning and embodied interaction, leveraging AR to connect formal mathematical concepts to perceptual and spatial reasoning.

As a proof of concept, the student will design and implement an AR prototype that allows users to interactively construct and manipulate sets and probabilistic scenarios (e.g., through visual regions, objects, or spatial zones) and receive immediate feedback. The system will be evaluated in a controlled user study to assess learning outcomes, conceptual understanding, and usability compared to more traditional representations.

In cooperation with Vittoria Frau (AU)

Kayaking in Virtual Reality: Designing an Interactive Testbed for Systematic Investigation of Cybersickness

This thesis focuses on the design, implementation, and evaluation of an engaging Virtual Reality (VR) kayaking experience that serves a dual purpose: providing an enjoyable interactive scenario while functioning as a controlled experimental platform for the systematic investigation of cybersickness.

Kayaking scenarios are particularly suitable for studying cybersickness. They combine visually rich motion (water flow, environmental movement, forward translation, and turns) with potentially strong visuo-vestibular conflicts. Such conditions frequently provoke discomfort, especially when interaction and active control are involved. While many existing studies rely on passive motion or highly constrained setups, this project aims to introduce meaningful user interaction into a controlled sickness-induction paradigm.

The work builds upon the previously developed and validated Cybersicker [1, 2], an experimental tool designed for highly systematic induction and measurement of cybersickness. While Cybersicker allows precise manipulation of motion parameters, it intentionally minimizes interaction. The proposed thesis extends this research line by creating a more interactive, ecologically valid VR application that maintains experimental controllability while enabling richer user engagement.

The resulting system will function as a reusable research platform for future studies on cybersickness, locomotion, interaction design, and mitigation techniques.

[1] Daniel Zielasko and Yuen C. Law. 2024. Cybersicker: An Open Source VR Sickness Testbed - Do you still have fun, or are you already sick? In Proceedings of the 30th ACM Symposium on Virtual Reality Software and Technology (VRST '24). Association for Computing Machinery, New York, NY, USA, Article 91, 1–2. https://doi.org/10.1145/3641825.3689503
[2] Zielasko, D., Rehling, B., von Dawans, B. et al. Do not immerse and drive? Prolonged effects of cybersickness on physiological stress markers and cognitive performance. Virtual Reality (2026). https://doi.org/10.1007/s10055-026-01335-0

Amplified Negative Gains in Redirected Walking: Designing and Evaluating Resistance-Based Locomotion Illusions in Virtual Reality

Redirected Walking (RDW) [1] is a locomotion technique in Virtual Reality that subtly manipulates the mapping between physical and virtual movement. By applying translation or rotation gains, users can physically walk within a limited tracked space while experiencing a much larger virtual environment. For example, a user may physically walk one meter while covering two meters in VR, enabling spatial compression without explicit awareness.

Traditionally, RDW is used to expand perceived virtual space and guide users through large environments within small physical areas. This thesis explores an alternative use case: intentionally applying negative or reduced translation gains to amplify the perception of resistance. Instead of making environments feel larger, the goal is to create the illusion of difficulty—such as walking through mud, wading through water, moving against strong wind, or being physically “held back.”

By systematically manipulating translation gains below unity, the project investigates whether users perceive increased effort, resistance, or environmental density—even though no physical resistance is present. The work examines perceptual thresholds, detection rates, and experiential consequences of such manipulations.

[1] N. C. Nilsson et al., "15 Years of Research on Redirected Walking in Immersive Virtual Environments," in IEEE Computer Graphics and Applications, vol. 38, no. 2, pp. 44-56, Mar./Apr. 2018, doi: 10.1109/MCG.2018.111125628.