Social navigation restrictions for interactive robots using augmented reality

Abstract

This paper describes the navigation mechanisms proposed for a mobile robot that uses augmented reality as interaction mechanism and laser scanners as main sensors. The peculiarities imposed by this interaction mechanism require continuous tracking of the person being escorted. The mechanism proposed for detecting and tracking people is based on a population of Kalman Filters and a basic association algorithm that matches past and new observations. The navigation system has been designed taking into account the special needs that an augmented reality interaction system imposes to a social navigation algorithm. This navigation system is integrated into a motivational control architecture that is also briefly described, as well as some preliminary experiments.