Evaluating the Influence of Character Realism on Avoidance Strategies in VR and a new method to generate virtual crowds using VR and motion capture
Picard, Stéven (2020)
Picard, Stéven
Åbo Akademi
2020
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe2020072747662
https://urn.fi/URN:NBN:fi-fe2020072747662
Tiivistelmä
The MimeTIC team of the IRISA research lab has a long history of studying human behaviours in order to create more natural virtual humans. Multiple studies were done in order to understand how pedestrians avoid each other, whose insights were used to develop new crowd simulation models. Since studying these behaviours is extremely complex, conditions in experiments are usually simplified to interactions between two walkers and this fact leads to a general lack of knowledge about these avoidance strategies within ecological situations. One of the particularities of the work done in the MimeTIC team is to leverage the use of Virtual Reality in order to explore more complex situations in a fully controlled experimental environment (for instance we should be able to simulate a crowd and see how an individual might react). In particular, VR was used to evaluate locomotion interfaces, as well as individuals avoidance of groups. However, a problem arises since it is known that this technology also creates other experimental limitations in comparison to the same real situations, which can potentially affect our decisions (e.g. we tend to underestimate distances in VR). Therefore, such limitations may also affect user locomotion behaviours in the presence of real and virtual obstacles, hence the knowledge thereafter used to develop new crowd simulation models.
That is why the MimeTIC team of IRISA is interested in exploring the influence of Character Realism on Avoidance Strategies in VR, in particular in relation to the realism of the displayed motions, trajectories, and visual representation. The technical objective of this master's thesis is to replicate an experiment previously conducted in a real situation, but where both participants are immersed in the same virtual environment (VE).
This first work has two goals:
- Understanding the effect of immersing two persons in the same virtual environment on their collision avoidance strategies, compared to real situations.
- Evaluating the influence of the degree of realism of the other person to interact with (in terms of body animation and global trajectories) on collision avoidance strategies.
Since during this thesis work it wasn't possible to complete this first study because of the COVID-19 crisis, another one is also described in this thesis. Some real life situations the MimeTIC team needs to analyse in VR might involve crowds, hence the need for realistic virtual crowds. However, creating these virtual crowds can be tedious, especially if we want those crowds to be realistic. This is why in this thesis we will try to see how the crowds generated in VR with an original method involving only one user and motion capture can be compared to crowds observed in real life experiments.
That is why the MimeTIC team of IRISA is interested in exploring the influence of Character Realism on Avoidance Strategies in VR, in particular in relation to the realism of the displayed motions, trajectories, and visual representation. The technical objective of this master's thesis is to replicate an experiment previously conducted in a real situation, but where both participants are immersed in the same virtual environment (VE).
This first work has two goals:
- Understanding the effect of immersing two persons in the same virtual environment on their collision avoidance strategies, compared to real situations.
- Evaluating the influence of the degree of realism of the other person to interact with (in terms of body animation and global trajectories) on collision avoidance strategies.
Since during this thesis work it wasn't possible to complete this first study because of the COVID-19 crisis, another one is also described in this thesis. Some real life situations the MimeTIC team needs to analyse in VR might involve crowds, hence the need for realistic virtual crowds. However, creating these virtual crowds can be tedious, especially if we want those crowds to be realistic. This is why in this thesis we will try to see how the crowds generated in VR with an original method involving only one user and motion capture can be compared to crowds observed in real life experiments.