We develop several use cases to demonstrate various Human Robot Interaction scenarios. In the first project, the task is to play football with a human. To do so, we implement a color detection algorithm to detect a ball in a scene. Then, the robot Aibo goes toward it and send it to the human. We explored a use case of Human-Robot Interaction where both human and robot have to interact together to play a game. We observed participants reactions to understand how to better design Natural Human-Robot Interaction.
In a second use case, the task is to put a robot, here Nao, in an environment and to teach him using Reinforcement Learning rules to interact with humans. Reacting to a stimulus, Nao is performing an action that humans encourage or not. After several iterations, Nao can almost interact properly with humans.
We explored the design space with various stimulus and different contexts such as outdoor and indoor