A voir également
- Short Bio
- Publications -----------------------------
- Work during Ph.D.
Work during Ph.D.
Title : Doctorant.e
No longer in the unit
My work mainly focuses in learning behaviors to interact with people. In order to do this, I first concentrated in group modelling, since I wanted a robot to interact not only with one person, but with multiple people (a group of people). Then, given a representation of this, work with some learning algorithms in order to approach a person or a group of people and in the future to guide a group people in an airport scenario, based on the SPENCER project scenario, which is the robot interacting with people at an airport.
Group Modelling
This part is important in order to have an understanding of the definition of a group of people, so a robot can know when if it's or not included in a group of people. Also we plan to use this information in the future as heuristics for approaching people. We have developped 2 approaches, a static one and a dynamic one. The static is based on the Marco Cristani's work, nonetheless when trying to use it in a dynamic environment we failed to detect groups due to the lack of dynamic variables so we created a formalism in the dynamic case.
Static Environment
Dynamic Environment
Formalism based on relative position and velocity between people in order to create a link between them, and then another formalism to designate the groups. Tested in a real scenario and a simulated scenario.
Learning To Approach People
Using IRL and (by now) a simulated environment in order to create navigation planners that take into account the best way to approach people in an optimal learning perspective.
Scenario with one person
With a human centered map, which is then mapped in an MDP, we give demonstrations to a robot on how it's suppose to get close to a person. Then we apply IRL (based on Sergey Levine's Work), and with the results we create a path, that can be considered as a global planner path in the navigation stack.
Multiple People
This is work in progress.