Master thesis proposals spring 2027
IoT / Mobile Development / Wireless Communication
Connecting machines: How can we make mobile-to-machine communication faster, smarter, and more reliable?
As industrial machines become increasingly connected, mobile applications play an important role in monitoring, controlling, and interacting with them. However, establishing fast, reliable, and seamless communication between smartphones and industrial equipment comes with several technical challenges.
Different communication technologies, such as Bluetooth Classic, Bluetooth Low Energy (BLE), Wi-Fi, and USB, offer different advantages and limitations in terms of performance, reliability, compatibility, and user experience. Choosing the right technology is only part of the challenge. How data is transmitted, how connections are established and maintained, and how different mobile platforms handle communication can significantly affect the overall experience.
At THNX, we develop THNX Connect, a solution that enables mobile applications to communicate with industrial machines using technologies such as Bluetooth and BLE, often involving CAN-based communication protocols.
In this master thesis, you will explore how communication between mobile devices and industrial machines can be improved, from the underlying communication technology and data transmission to connection management and user experience.
Possible areas to explore
Communication technologies and performance
- Investigate and compare Bluetooth Classic, BLE, Wi-Fi, and USB for communication between smartphones and industrial machines.
- Evaluate performance characteristics such as latency, throughput, reliability, and energy consumption.
- Explore differences between Android and iOS and how platform limitations influence the choice of technology.
- Investigate how local Wi-Fi communication with a machine can coexist with the smartphone's regular internet connectivity.
Optimizing industrial data transmission
- Investigate how industrial communication protocols, such as CAN and CANopen, can be transmitted efficiently over wireless connections.
- Explore how BLE services and characteristics can be designed to improve communication performance.
- Experiment with techniques such as buffering, batching, message prioritization, and data compression.
- Develop tools or prototypes for measuring and analyzing data transmission, potentially using existing CAN analysis tools and real machine data.
- Identify bottlenecks and evaluate strategies for improving responsiveness and reliability.
Reliable connectivity and user experience
- Investigate how connection establishment, pairing, authentication, and reconnection can be improved.
- Explore platform-specific challenges when communicating with external devices on Android and iOS.
- Evaluate how applications can handle interrupted connections, device switching, and authentication issues.
- Investigate ways to make connecting to industrial machines more seamless and intuitive for end users.
What you will do
- Investigate existing research, technologies, and best practices for mobile-to-machine communication.
- Identify and formulate relevant research questions within one or more of the areas above.
- Design and implement proof-of-concept solutions or experimental test environments.
- Conduct practical experiments to compare communication strategies and measure their performance.
- Evaluate results and identify trade-offs between performance, reliability, compatibility, and user experience.
Goal
The goal of this master thesis is to investigate and demonstrate how communication between mobile applications and industrial machines can be improved.
The expected outcome is a technical evaluation, supported by practical experiments and one or more proof-of-concept implementations, that provides insights and recommendations for developing efficient, reliable, and user-friendly mobile-to-machine communication solutions.
Depending on the chosen focus, the project could result in an optimized BLE communication implementation, a comparison of alternative communication technologies, an improved connection management strategy, or a framework for measuring and evaluating industrial communication performance.
AI / IoT / Machine Learning / Industrial Communication
From unknown machine to connected application: Can AI understand CAN data?
Connecting a new machine to an existing digital platform often requires significant manual effort to understand its communication protocols, identify relevant messages, and map them to a standardized format. But what if AI could help automate this process?
At THNX, we develop THNX Connect, a solution that enables mobile applications to communicate with industrial machines. Integrating machines from new customers can involve analyzing CAN communication, identifying data structures, and creating custom adapters to translate machine-specific messages into formats supported by THNX Connect.
In this master thesis, you will investigate whether AI can analyze CAN communication from previously unknown machines and assist in mapping their messages to standardized protocols such as CANopen.
What you will do
- Investigate CAN and CANopen communication protocols and common message structures.
- Explore AI-based and rule-based techniques for identifying patterns, relationships, and potential meanings in CAN data streams.
- Investigate how recorded or live CAN traffic can be analyzed and matched against known message definitions and communication standards.
- Develop a proof of concept that analyzes CAN messages and suggests mappings to formats supported by THNX Connect.
- Explore whether AI can assist in generating communication adapters for customer-specific CAN implementations.
- Evaluate mapping accuracy, limitations, and the need for human verification.
Goal
The goal is to investigate how AI can simplify and streamline the work of integrating new industrial machines into THNX Connect.
The expected outcome is a proof of concept capable of analyzing CAN communication and suggesting message mappings or adapter configurations. The work should also include an evaluation of the solution's accuracy, reliability, and potential to reduce manual integration work.
Game Development / Augmented Reality / Technology Evaluation
Beyond Unity: Which technology should power the next generation of AR games?
Unity has long been a popular choice for developing games and augmented reality (AR) experiences. But are there alternative game engines that could offer a more efficient, flexible, or sustainable development process in the long run?
At Exenze, we create interactive, location-based experiences where the digital and physical worlds meet. Today we use technologies such as Unity and Vuforia, but we are curious about which other technologies could be relevant in the future.
In this master thesis, you will explore modern game engines and development frameworks and evaluate how well they are suited for creating location-based AR games.
What you will do
- Investigate alternative game engines and frameworks, such as Godot and other relevant technologies.
- Compare AR support, mobile platform support, performance, and developer experience.
- Explore how existing functionality built on Unity and Vuforia could be implemented using other technologies.
- Develop a proof of concept of a selected AR game experience using an alternative technology.
- Evaluate technical limitations, long-term maintainability, development costs, and opportunities for further development.
Goal
The goal is to investigate which game engines and technologies are best suited for the future development of location-based AR games.
The expected outcome is a technical evaluation and a proof of concept that together provide recommendations on future technology choices for Exenze's AR experiences.
Augmented Reality / Computer Vision / Mobile Development
AR without physical markers: Can virtual objects find their place in the real world?
Many AR experiences rely on physical markers to determine where virtual objects should be placed in the real world. But what if we could place characters, objects, and interactive experiences outdoors without using any physical markers at all?
Exenze's AR experiences currently use marker-based technology, among other techniques, to position virtual objects. We want to explore replacing or complementing this with modern positioning and tracking technology.
In this master thesis, you will investigate how virtual objects can be placed and tracked in outdoor environments using smartphone sensors and modern AR technologies.
What you will do
- Investigate techniques for markerless AR, visual positioning, and sensor-based tracking.
- Explore how GPS, smartphone sensors, and computer vision can be combined to improve positioning.
- Develop a prototype in which virtual characters or objects are placed at predefined outdoor locations.
- Investigate how virtual characters can follow defined movement patterns or paths, or move between different positions.
- Evaluate positioning accuracy, stability, and user experience in different outdoor environments.
Goal
The goal is to investigate the possibilities of creating outdoor AR experiences without depending on physical markers.
The expected outcome is a proof of concept demonstrating markerless AR in an outdoor environment, together with an evaluation of the technology's possibilities and limitations for location-based games.
Mobile Development / Geolocation / Cross-Platform
How accurate is accurate enough? GPS positioning in mobile and web applications
Location-based apps depend on knowing where the user is. But how much does positioning accuracy actually differ between a web app, a native app, and an app built with a cross-platform framework?
When we create outdoor games and interactive experiences, a difference of a few meters in positioning can be decisive for how the user experiences the game. At the same time, different platforms and development technologies offer different ways of accessing the smartphone's positioning capabilities.
In this master thesis, you will investigate and compare how different mobile development technologies handle geolocation, with a focus on accuracy, performance, and user experience in real-world environments.
What you will do
- Investigate the possibilities and limitations of geolocation in web apps, native apps, and cross-platform frameworks.
- Compare selected technologies, such as React Native, Kotlin Multiplatform, and native iOS or Android development.
- Develop a test application or framework for collecting and analyzing location data.
- Conduct practical tests to measure positioning accuracy, update frequency, stability, and battery consumption.
- Explore how positioning uncertainty can be managed through technical solutions or smart game design.
Goal
The goal is to investigate how the choice of development technology affects geolocation in mobile applications.
The expected outcome is a comparative evaluation with practical measurement results and recommendations on which technologies are best suited for location-based games and interactive outdoor experiences.
Game Design / Gamification / UX / Location-Based Experiences
Turn the world into a game board: How can we create more fun and engaging outdoor games?
What makes people want to go outside, explore new places, and keep playing? Location-based games combine physical activity with digital experiences and make it possible to turn everyday environments into exciting game worlds.
But how can we make such experiences even more fun, engaging, and motivating?
At Exenze, we work with location-based games and experiences, including Run and Find. We see opportunities to develop the concept further with new game mechanics, mini-games, and ways of combining the digital game world with the physical surroundings.
In this master thesis, you will explore how gamification, game design, and interaction design can be used to create new kinds of outdoor games that motivate users to get moving, explore their surroundings, and play together.
What you will do
- Investigate successful game mechanics in location-based games, treasure hunts, orienteering, and other outdoor activities.
- Explore how progression, rewards, competition, cooperation, and mini-games affect motivation and engagement.
- Develop and evaluate concepts for new game modes, mini-games, or interactive experiences.
- Design and develop a playable prototype of a selected concept.
- Conduct user tests to study enjoyment, usability, engagement, and motivation.
Goal
The goal is to investigate how game mechanics and gamification can be used to make location-based outdoor experiences more engaging and motivating.
The expected outcome is an evaluated game concept and a playable prototype demonstrating how new ideas can shape the future of location-based games.
Artificial Intelligence / RAG / Evaluation
Can we trust AI? How do we build more accurate and reliable AI assistants?
AI-based support assistants can quickly answer complex questions, but how do we know the answers are actually correct? How do we make sure an AI assistant relies on trustworthy information instead of making up its own answers?
More and more companies use AI to help their customers find information and solve problems. At the same time, this places high demands on the quality of the generated answers. An AI assistant that sounds convincing but gives incorrect information can create more problems than it solves.
In this master thesis, you will investigate how techniques such as Retrieval-Augmented Generation (RAG) and grounding can be used to improve the accuracy and reliability of AI-based support assistants.
What you will do
- Investigate different approaches to RAG and grounding, where the AI model's answers are anchored in relevant and reliable information.
- Explore strategies for finding, selecting, and presenting the right information to a language model.
- Compare methods for reducing hallucinations and improving the relevance and correctness of answers.
- Develop a framework for evaluating AI-generated answers using realistic support questions.
- Experiment with different implementations and configurations to measure answer quality, accuracy, and reliability.
Goal
The goal is to investigate how we can build AI assistants that give more correct, relevant, and reliable answers.
The expected outcome is an evaluation framework and a proof of concept demonstrating how different techniques can improve the quality of AI-based support assistants.
Artificial Intelligence / Knowledge Management / UX
A knowledge base that keeps itself up to date: Can AI simplify the future of support?
An AI assistant is never better than the information it has access to. But keeping a knowledge base correct, complete, and up to date can be both time-consuming and complicated, especially when the information is spread across documents, systems, and different sources.
What if AI could not only use the information in a knowledge base, but also help us improve and maintain it?
In this master thesis, you will explore how AI can be used to simplify knowledge management and make it easier for both technical and non-technical users to quality-assure and further develop AI-based support solutions.
What you will do
- Investigate the challenges of creating, structuring, and maintaining knowledge bases for AI-based support systems.
- Explore how AI can identify information that is missing, outdated, duplicated, or contradictory.
- Investigate how AI-generated answers can be evaluated and used to identify gaps in the underlying knowledge.
- Design and develop a proof of concept for a self-service interface where users can easily review, update, and improve the content of a knowledge base.
- Evaluate how the solution can make it easier for people without a technical background to maintain and quality-assure an AI assistant.
Goal
The goal is to investigate how AI can make the work of keeping knowledge bases up to date and relevant more efficient.
The expected outcome is a proof of concept for an AI-assisted workflow that simplifies knowledge management and helps improve the quality of answers from AI-based support assistants.
Agentic AI / Data Analysis / Automation
Beyond the chatbot: Can autonomous AI agents turn data into valuable insights?
AI is no longer just about asking questions and getting answers. Autonomous AI agents open up new possibilities to continuously monitor information, identify patterns, detect anomalies, and suggest actions, without a user having to initiate every single analysis.
But how useful and reliable are these agents in practice? And how can we build solutions that work independently without compromising security or control?
In this master thesis, you will investigate how agentic AI can be used to analyze information from digital services, documents, and connected systems, with the aim of turning large amounts of information into useful insights.
What you will do
- Investigate frameworks and tools for agentic AI, such as OpenClaw or other relevant alternatives.
- Explore techniques for collecting, combining, and analyzing information from multiple data sources.
- Develop a proof of concept in which an AI agent monitors selected digital services or analyzes business information.
- Investigate how AI agents can detect anomalies, identify trends, and generate relevant recommendations.
- Evaluate the reliability of the agents and the quality of the insights they generate.
- Investigate challenges related to security, privacy, control, and autonomous decision-making.
Goal
The goal is to investigate how autonomous AI agents can be used to turn information from digital services and systems into concrete, useful insights.
The expected outcome is a proof of concept demonstrating the possibilities of agentic AI, together with an evaluation of the technology's limitations and recommendations for how it can be used safely and effectively.