An on-line calibrated visual intelligent space for navigation and control of mobile robots

Abstract

An intelligent space employing multiple vision sensors is proposed. The union of the sensing ranges of these vision sensors covers a large unknown workspace where a mobile robot can proceed to explore. The goal of this research is to accomplish mobile robot localization and navigation based on such an on-line calibrated sensor network. In particular, each adjacent IP camera pair is assumed to have an overlapping field of view, but the positions and orientations of the IP cameras with respect to either the Cartesian space or any mobile robot are unknown. The idea is to control the mobile robot to actively perform calibration with any stationary IP camera that can observe pre-selected features onboard the mobile robot, and thus establish coordinate transformations among these IP cameras. The proposed system only requires a PC and a network of stationary IP cameras without performing off-line calibration. Therefore, this seemingly novel approach appears to be low-cost, effective, efficient, and flexible.