Avatar VR Bachelor Project
#Unity Development
#research
#User survey
The aim of the project was to investigate the effects of avatars with different degrees of reality in a virtual reality learning environment on learners. To this end, we designed and developed an immersive virtual reality application that acted as a test environment. The learning application was then tested on users, who subsequently completed various questionnaires. When evaluating the questionnaires, we were then able to investigate whether and to what extent the different visualization of the character influenced the relationship and perception of the user.
TEAM
Robert Dolibog
Vladyslav Kyselov
Imanuel Leiserowitz
Nathalie Claire Huppert
Karolina Serova
Danylo Vovochok
PROGRAMS
Concept
The creation of the testing environment involved the construction of an immersive VR application, between 5 to 10 minutes of Playtime, designed to investigate the effects of different avatars on the learner. In this application, an avatar introduces itself and provides a short instructional session, enabling the study of various impact factors. In our learning environment, the focus was on the 5 safety rules of electrical engineering. To make this instructional session as hands-on as possible, we partnered with the company se.services. The local electronics service provider, was a helpful partner in providing us with priceless expertise in the field of employee training.
In collaboration with their Head Instructor Markus Zippel, who has trained countless of electricians during his career, we were able to create a scenario, deliviring a basis on which we could later evaluate the users experiences.
Phase 1: The literature research
During our process, we intensively examined over 30 scientific papers and studies. This comprehensive analysis formed the basis for delving deeper into the topic and developing our own questionnaire concept. In addition, the findings of various papers helped us to design the learning environment and, above all, the avatars appropriately.
From all these studies, we selected 3 suitable papers that already contained questionnaires and created our own questionnaire based on them. We were able to query four categories:
Social Presence
telepresence
trust
affect
Statistical Investigation
For the statistical investigation of our project, we first created a research model. This represents several correlations, which we took from our literature research and adapted to our survey environment. In addition, we set up 4 hypotheses which we tried to prove or disprove at the end.
Hypotheses:
1. A higher degree of realism of avatars compared to cartoon avatars leads to them achieving higher social presence scores among users.
2. Users who feel a higher affect towards an avatar also have a higher level of trust in this avatar.
3. A higher degree of realism of the avatars has a positive effect on the perceived affect.
4. A high level of avatar realism has a positive effect on the level of trust felt for that avatar.
Phase 2: Design and Development
In our study, we wanted to explore possible variations in the player’s relationship with the instructor based on its visual representation. One group of participants interacted with a human-looking 3D model, while the other group received instructions from a more abstract representation, a robot.
One goal of our project was to construct a virtual test environment where the interaction between a user and the instructor could be examined. The challenge consisted of creating such an environment while not making it distracting enough to turn the focus away from the character that has to be evaluated. For a majority of our testers this was the first time experiencing a surrounding in Virtual Reality.
Scene 1: Introduction inside the Apartment
In this scene the player is greeted by the instructor, and asked to point out some different uses of electric current in a regular household. To learn about the transformation from electric current to heat, the player needs to interact with the toaster. To activate an electric magnet one needs to point out the Toy Crane, and by using the lamp, the player can light up the scene. Supported by some amusing voicelines the player is playfully introduced into the topic of electric energy, before being confronted with more complex topics.
Scene 2: Lesson in the Workshop
In this workshop setting, the avatar guides the user through the differences between a basic circuit diagram, as commonly known from physics classes, and an electrician’s circuit diagram. Not only do the symbols differ, but now there are also ciruits switches included in the drawings. The instructor emphasizes a fundamental rule in the electrician’s craft: always disable the circuit switches before working on the electrical components!
The instructor then challenges the user to reconstruct the circuit. Various 3D assets, including a switch, a lamp, and cables, are laid out on the table in front of the player. A template next to them allows the components to be placed and snapped into place. The power source and fuse switches are already mounted on the edge of the template, animated to signify electrical flow.
Ideally, the user starts building without deactivating the fuse switches. If a component under power is touched, a loud bang surprises the player, and the environment disappears. The instructor’s voice from offscreen explains that the first and most important safety rule has been overlooked.
Now, the user has the opportunity to try again, earning the instructor’s praise once they successfully build the circuit without causing an explosion.
Motion Capturing
To capture the instructors movements we recorded the speech of the Head Instructor of the electronics company aswell as his movements, in the Lab of the creative media team of the “Institut für Kultur und Informatik an der HTW Berlin” which draws a rigged skeleton by capturing the movements from 8 angles.
- Captury (for real-time motion-capturing)
- Blender (for cutting and cleaning the captured data)
- Logic (Digital Audio Workstation for cleaning up audio clips)
Phase 3: User testing
In order to examine if the difference in character design had an impact on the relationship with the teacher, we conducted tests on two groups of participants during the span of two days on campus. The participants filled out questionnaires to provide insights into their experience and perception.
Phase 4: Test results and analysis
To analyze the test results, we employed statistical methods in SPSS, specifically ANOVA and linear regression, after cleaning the data. The hypotheses related to social presence and trust could not be confirmed, possibly due to the distracting nature of the open environment or the lack of personal connection to the avatars. Future research should explore how allowing users to select their avatars might enhance connection and social presence, as well as examine the impact of user agency on these experiences. Such studies could help create virtual environments better tailored to user needs.
Phase 5: Showtime
At the end of the semester there was a big showtime where all groups presented their projects and the professors could ask questions, this made up 30% of the final grade of the project. In addition, many students came and tried out and examined the various projects. In addition, each group created a website for their own project.
The showtime was divided into three parts according to which the projects were graded:
Presentation
Booth
Answering the questions
Final grade for the entire project:
1,0
Posters for our booth
Conclusion
However, with team support and advice from our supervisors, we found suitable papers. Our teamwork improved over the semester, and I developed new skills, including programming in C# and working with Unity. Overall, this project was a great experience, and I believe the knowledge gained will benefit me in future endeavors.