Objective:

• Develop and deploy a web-based decision-support platform for the early detection of language disorders in children aged 0 to 5 years, ensuring it is ready for use in a real clinical environment.

Key Components:

• Develop the FrontEnd of the web platform using modern technologies such as React.js.
• Design the user interface following UX principles.
• Use appropriate frontend frameworks while considering the already deployed backend technology.
• Basic knowledge of graph-based databases.

Challenges:

• Integrating previous development into a fully functional and production-ready system.
• Ensuring compatibility with the existing backend infrastructure.
• Designing an accessible interface for different types of users.

Objective:

• Develop an AI-powered sign language translation system that converts sign language gestures into real-time subtitles or spoken words.

Key Components:

• Use Vision Pro’s/meta Quest hand tracking and AI-based gesture recognition to interpret sign language.
• Develop a real-time text and voice translation overlay for mixed-reality conversations.
• Integrate with speech-to-text AI models for two-way communication between signers and non-signers.
• Enable customization for different sign languages (ASL, BSL, etc.).

Challenges:

• Achieving high accuracy in gesture recognition across different signing speeds.
• Addressing regional sign language variations.
• Minimizing latency for real-time conversation flow.
• Document the outcomes in a scientific manuscript.

Objective:

• Design a spatial mind-mapping application that allows users to visualize, organize, and interact with their thoughts, tasks, and ideas in an immersive mixed-reality space.

Key Components:

• Develop a spatial UI allowing users to place and arrange concepts freely.
• Use gesture and gaze tracking to create, link, and categorize ideas.
• Implement AI-assisted organization, suggesting connections and clustering related concepts.

Challenges:

• Ensuring intuitive navigation in a three-dimensional space.
• Preventing information overload by optimizing visual organization.
• Balancing AI assistance with user control for mind mapping.
• Document the outcomes in an academic manuscript. 

Objective:

• Create an AI-powered personalized learning environment, enabling adaptive, interactive, and immersive educational experiences.

Key Components:

• Implement AI-driven learning models that analyze user engagement, progress, and preferences.
• Develop spatial learning environments where lessons adapt based on user interaction.
• Integrate 3D interactive elements (e.g., virtual chemistry labs, historical reconstructions).

Challenges:

• Creating engaging educational content optimized for mixed reality.
• Ensuring personalization without overwhelming the user with excessive data.
• Managing AI biases and accuracy in content recommendations.
• Propose the evaluation of long-term effectiveness compared to traditional learning methods.
• Document the outcomes in a research manuscript.

Objective:

• Investigate the effectiveness of gaze-based navigation in mixed reality applications, optimizing UI/UX for seamless hands-free interaction.

Key Components:

• Develop a gaze-based interface, enabling users to interact with elements using eye tracking.
• Compare gaze-based vs. traditional (gesture/touch) interactions in different contexts (e.g., gaming, productivity, social apps).
• Conduct usability testing to measure precision, fatigue, and user preference.
• Optimize dwell-time settings for natural interaction without excessive eye strain.

Challenges:

• Preventing accidental interactions due to involuntary eye movement.
• Balancing gaze responsiveness without causing discomfort or fatigue.
• Addressing accessibility concerns for users with vision impairments or eye-tracking difficulties.
• Ensuring smooth multi-modal integration with gestures and voice commands.
• Document the outcomes in an academic manuscript

Objective:

• Develop a real-time object recognition system for Apple Vision Pro that allows users to interact with real-world objects in mixed reality using AI-powered detection and gesture-based controls.

Key Components:

• Implement AI-based computer vision for object detection using Apple Vision Pro’s LiDAR and cameras.
• Develop gesture-based interactions, allowing users to manipulate digital overlays on recognized objects.
• Integrate object classification and contextual information, providing real-time insights (e.g., scanning books for summaries, identifying tools for usage tips).

Challenges:

• Ensuring low-latency processing for real-time object recognition.
• Optimizing AI models for efficient on-device performance.
• Handling varied lighting conditions and occlusions in real-world environments.
• Designing an intuitive gesture-based interaction system.
• Document the results in an academic manuscript

Objective:

• The goal of this thesis is to explore the application of Learning Agents in Unreal Engine 5.5, leveraging reinforcement learning techniques to create adaptive and intelligent non-player characters (NPCs). The study will focus on designing agents that dynamically adjust their behavior based on real-time interactions, improving game AI realism and decision-making.

Key Components:

• Understanding Learning Agents in Unreal Engine:
  • Review Unreal Engine’s Learning Agents framework.
  • Analyze reinforcement learning (RL) models and their implementation in UE5.5.
• Developing AI-Driven NPCs:
  • Create an environment within Unreal Engine where NPCs learn from player actions.
  • Implement decision-making algorithms using behavior trees and machine learning models.
• Training AI with Reinforcement Learning:
  • Apply reward-based learning for AI agents to improve navigation, combat, and interaction.
  • Test and evaluate different training methodologies (e.g., Q-learning, PPO, DDPG).
• Performance Evaluation and Optimization:
  • Compare the efficiency of trained agents versus traditional scripted AI.
  • Optimize training cycles to balance realism and computational efficiency.

Challenges:

• Computational Costs: Training complex AI agents in a real-time environment may require significant processing power.
• Balancing AI Behavior: Ensuring the NPCs learn efficiently while maintaining an engaging user experience.
• Integration with Game Mechanics: Aligning learning agents with existing UE5 mechanics like physics, animations, and interactions.
• Data Collection and Training Time: Reinforcement learning requires extensive training data, which can be time-consuming.

Objective:

• This thesis aims to create a comprehensive manual for collaborative VR development in Unreal Engine 5. The manual will provide best practices for content structuring, asset organization, coding standards, and version control, ensuring seamless teamwork in VR projects.

Key Components:

• Understanding Collaborative VR Development Needs:
  • Analyze common challenges faced by teams working on VR projects in UE5.
  • Identify bottlenecks in asset management, level design, and scripting.
• Implementation of UE5 Style Guide Principles:
  • Content Directory Structure: Organize assets efficiently for large-scale VR projects.
  • Naming Conventions: Ensure clarity and consistency across files, blueprints, and scripts.
  • Blueprint Best Practices: Standardize event-driven logic for modular and reusable VR interactions.
• Collaborative Workflow Optimization:
  • Establish version control strategies (Git) for team-based VR development.
  • Develop guidelines for merging assets, blueprints, and level design to prevent conflicts.
  • Create a pipeline for integrating VR assets, physics interactions, and animations smoothly.
• Manual Development and Case Study:
  • Compile findings into a structured manual for teams developing VR experiences in UE5.
  • Validate the guide by implementing it in a small-scale collaborative VR project.
  • Gather feedback from developers and refine the manual based on practical usability.

Challenges:

• Standardizing VR-Specific Development Practices: Unreal Engine’s general guidelines may need adaptation for VR projects.
• Managing Large-Scale Collaboration: Ensuring seamless workflow among designers, programmers, and artists.
• Avoiding File Conflicts in Version Control: Strategies to prevent merge issues in blueprints, assets, and VR-specific shaders.
• Ensuring Scalability: Creating a guide applicable to both small indie teams and larger VR production studios.

Objective:

• Investigate best practices for designing user interfaces specifically for VR applications and implement a UI framework.

Key Components:

• Research existing UI design principles for VR.
• Develop a VR UI framework that includes common elements like menus, buttons, and interaction methods.
• Create a sample VR application to demonstrate the framework and conduct usability testing.

Challenges:

• Ensuring the UI is intuitive and easy to use in a 3D space.
• Addressing potential issues like motion sickness and interaction accuracy.

Objective:

• Investigate and implement advanced rendering techniques to improve the performance and visual quality of VR applications.

Key Components:

• Explore techniques like foveated rendering, dynamic resolution scaling, and level of detail (LOD) management.
• Develop a VR prototype demonstrating these techniques.
• Conduct performance and visual quality evaluations to assess the improvements.

Challenges:

• Balancing performance gains with visual fidelity.
• Ensuring the techniques are compatible with various VR hardware.

Objective:

• Develop AI algorithms that dynamically adapt VR environments based on user interactions and preferences.

Key Components:

• Create AI models that learn from user behavior and adjust the VR environment in real-time.
• Develop a VR application that incorporates these adaptive environments.
• Evaluate the impact of adaptive environments on user engagement and experience.

Challenges:

• Designing AI models that can operate in real-time with minimal latency.
• Ensuring the adaptive changes enhance the user experience without causing discomfort.

Bachelor thesis

_{Bachelor thesis}

Master thesis

↓ Winter 2024/25

• Practical Course: Heilbronn: Fundamentals of Programming

↓ Summer 2024

• Lecture: Heilbronn: Introduction to Software Engineering

↓ Winter 2023/24

• Lecture: Heilbronn: Introduction to Informatics
• Practical Course: Heilbronn: Fundamentals of Programming

↓ Summer 2023

• Lecture: Heilbronn: Introduction to Software Engineering

↓ Winter 2022/23

• Lecture: Heilbronn: Introduction to Informatics
• Practical Course: Heilbronn: Fundamentals of Programming
• Seminar: Interactive Learning
• Seminar: Heilbronn: Interactive Learning

↓ Summer 2022

• Lecture: Heilbronn: Introduction to Software Engineering

↓ Winter 2021/22

• Lecture: Heilbronn: Introduction to Informatics
• Practical Course: Heilbronn: Fundamentals of Programming