Camera Aided Inertial Navigation in Poor GPS Environments

Abstract

This paper presents a framework for robust localization suitable to unmanned systems operating in environments that have limited or unreliable access to external localization information, such as GPS. The core algorithm presented is a recasting of the Simultaneous Localization and Mapping (SLAM) technique from the robotics literature as an aid to inertial navigation. A single camera is used to track unknown environment features providing a consistent constraint on inertial sensor drift while global localization information is unavailable. Results from an unmanned flight vehicle, performing a tracking and surveillance task with unreliable GPS are presented.