Multi-agent system for fast deployment of a guide robot in unknown environments

Nowadays, deploying service robots and adapting their services to a new environment is a task which might require several days. This is an important problem of robotics in general, but specially when the goal is to bring robots to our everyday's life. In this paper we present a multi-agent intelligent space, which consists on intelligent cameras and autonomous guide robots. The deployment of the system does not require expertise and can be done in a short period of time. The cameras detect situations requiring the robots' guiding services, inform the robots accordingly, and support the robots navigation towards the goal areas, without the need of a map of the environment. An example of these situations requiring the robot guide service could be a group of persons entering a museum. In this sense, we also present and adaptive person follower behaviour intended to be the basis of a route learning process, necessary to offer the guide service.

keywords: Guide robot, intelligent space, person follower, feature weighting, image retrieval.