Human-robot dialogue and collaboration for social navigation.

The navigation of a robot in an environment with humans is a subject with enormous interest in the last years. To be accepted in these types of scenarios, it is important that the robot navigate respecting social norms, for example, avoiding getting too close to humans, avoiding interrupting a conversation or asking permission to pass through a blocked path. This project aims to describe the dialogue manager, besides the corpus that allows establishing dialogues between the robot and the humans in real situations to improve the behavior of the robot navigation system, making it more socially accepted. The dialogue manager will be a RoboComp agent, which reads information from the robot’s world representation (a graph) and adapts the conversation to the current situation (e.g, according to the age of people, genre, etc).