Personality of Virtual Coaches
Personality: Gestures and Poses
Project Team
|
Supervision
|
Introduction
In our project “Personality: Gestures and Poses” we developed a personal training environment, where you can choose the personality of your trainer. Depending on the personality of the trainer, their appearance and movements during a workout routine change. This should help a wide variety of trainees to always feel supported during a training exercise and moreover get perfectly motivated to exercise.
Design
The scene of the program is divided into three parts. On the left side, we implemented the UI for choosing how many repetitions and exercises one wants to do. In the middle of the scene, we show the animated character and on the right side, before starting with the training you have a button list of all animations the character can show.
Implementation
In our final application the user can choose between 5 different trainers. Each one of them has its own individual personality and reacts individually to the input of the user. When switching through the characters, each character moves uniquely depending on the personality. While implementing the animations we stuck to the paper “Relating Personality and Behavior: Posture and Gestures” by Ball and Breese, published in 1999. Ball and Breese created a Bayesian network of movements and gestures depending on certain personality traits. For example a friendly personality as well as positive emotional valence is communicated by postures and gestures that increase accessibility to a conversational partner. After choosing the trainer and the amount of exercises and repetitions the user can start the training. Every trainer will react with a unique animation and the timer for the training begins. If the user takes a certain amount of time for the exercise, the trainer will react with a gesture.
Future Work
For future work it would make sense to create our very own animations to complement the personalities and empower their effectiveness to motivate and emphasize. In addition, adding not only speech bubbles, but getting the trainers to talk to the user would be a huge step to humanizing the models.