This project focuses on the integration of sensor information and linguistic descriptions in order to guide the navigation of a robot that works in previously unknown environments. Route and office descriptions provided by persons that know the working area will be used to obtain general information of the kind of objects to be found and their global distribution in the environment. Ultrasonic sensors will provide metrical information of the environment outline, while vision will be used for object identification in order to solve conflicts, confirm expectations and focus on areas of interest. In this way, we will use three independent sources of information, each one of different nature and characteristics, but complementary among them in such a way that from their integration we can obtain the necessary information to solve the navigation tasks. We will also consider the problems of how to represent and propagate (using fuzzy sets) the uncertainty that is inherent both to the linguistic descriptions and to the perceptual information. Particular importance will be given to experimentation with the mobile robot already available at the host institution.
This project proposal combines fields of research where the applicant has already a long experience with other fields that are new to him and where he can strongly benefit from the knowledge and expertise form the members of the host institution. The training he will acquire is also of significant interest for his research centre at his country of origin in order to continue this research line at the end of this project.