Improving the Human-centered Robot Navigation Agent

My proposal consists on transform the fixed personal space used actually in Robocomp into an adaptive space, depending of the spatial context (there is difference, for instance, between a narrow corridor and a room). The main idea is to change the values that defines the personal space function in order to adapt the personal space to allow the robot to navigate around the person without problems, adapting it to the spatial context. Besides, I would like to extend the social navigation agent, including other typical social behaviors: crossing people in corridors, approaching to humans in human-robot interaction, etc. In this respect, it would be interesting to study how these algorithms are also integrated into the planner and mission agents, and how define the priority between these missions by defining a social behavior planner.