Description

The main objective of the presented thesis will be to develop new techniques in Human-Robot collaboration as partners in accompaniment and approaching tasks.

Several new methods of human-robot collaboration have been developed. Three of them in the field of accompanying people by robots. The first of these allows the robot to accompany a person using an adaptive side-by-side accompaniment to the environment, and the remaining two allow the robot to accompany a group of two people through two different formations: side-by-side and training in V. A method of approaching people in a natural way has also been developed using trajectories generated by polynomial curves. Finally, a new method was developed that combined both tasks, the accompaniment and the approach, with which the robot interacts with two people, the one accompanying and the one approaching.

To obtain all these methods, we have reformulated the Extended Social Force Model many times, and we have extended the cost functions to select the best robot’s navigation path which optimizes the implemented task. Furthermore, we have developed several metrics to evaluate the robot’s behavior in all the methods, in terms of quantitative evaluation and qualitative evaluation through surveys that obtain the thoughts of people about the interaction with the robot.

In addition, a new method of interaction had to be developed through a game and various communications of current states of the robot, to achieve a more natural accompaniment by inexperienced people in interactions with robots. Also, a methodology to conduct the real-life experiments with people non-expert in robotics is included in this thesis.

The thesis is part of the National project (Interacción, aprendizaje y cooperación robot humano en áreas urbanas) which aim is to go a step further in the deployment of the social robots in urban environments. This framework will be used in the tests of the research results and we will improve the robot abilities when is collaborating with humans.

The work is under the scope of the following projects:

• Robot-Int-Coop: Robot-Human Interaction, Cooperation and Learning in Urban Areas (web)
• ColRobTransp: Colaboración robots-humanos para el transporte de productos en zonas urbanas (web)