Service robots must be often able to navigate in complex indoor environments. They may also sometimes be instructed by humans to go to specific places. If the robot does not already know how to go there, it will need to follow human instructions about the route to follow, like for instance:
"first go straight until you arrive at the end of the corridor, then turn left. Once you see a door, turn right, continue for 10 meters, and you have arrived"
The master thesis will investigate
- how to represent such route directions in a way that can be understood (and acted upon!) by the robot
- how the robot can ask follow-up questions to ensure it has correctly understood the directions, and clarify potential misunderstandings.
The approach will be implemented and evaluated on an actual Pepper robot.
If several students are interested in this topic and are willing to collaborate, this thesis may be adapted to accommodate the work of two students.
Prerequisites:
Good programming skills in Python, and ideally some knowledge of Java/Kotlin. Interest for robotics, human-robot interaction and spoken dialogue systems.