In order for unmanned vehicles to be able to successfully accomplish the planned task in high-risk rough-terrain situations, an increased level of autonomy is primarily required. In this respect, localization becomes a key aspect of onboard rover functionality. Usually, mobile robot missions have relied on odometric sensors such as wheel encoders and inertial measurement units/gyros for motion estimation. While these offer a simple solution, they are prone to wheel-slip in loose soil and drift of biases, respectively. Alternatively, the use of visual landmarks observed by stereo cameras to localize a rover offers a more robust solution but at the cost of increased complexity. This paper presents current work at the University of Salento in the development of a vision-based algorithm for accurate and robust 6-DoF ego-motion estimation. The details of the various steps of the proposed approach are described, and field trial results, obtained using an all-terrain rover, are presented proving the feasibility of this method.
A Visual Odometry Algorithm for Rough-Terrain Mobile Robots
DI CASTRI, CARMELO;GIANNOCCARO, NICOLA IVAN;MESSINA, Arcangelo;REINA, GIULIO
2008-01-01
Abstract
In order for unmanned vehicles to be able to successfully accomplish the planned task in high-risk rough-terrain situations, an increased level of autonomy is primarily required. In this respect, localization becomes a key aspect of onboard rover functionality. Usually, mobile robot missions have relied on odometric sensors such as wheel encoders and inertial measurement units/gyros for motion estimation. While these offer a simple solution, they are prone to wheel-slip in loose soil and drift of biases, respectively. Alternatively, the use of visual landmarks observed by stereo cameras to localize a rover offers a more robust solution but at the cost of increased complexity. This paper presents current work at the University of Salento in the development of a vision-based algorithm for accurate and robust 6-DoF ego-motion estimation. The details of the various steps of the proposed approach are described, and field trial results, obtained using an all-terrain rover, are presented proving the feasibility of this method.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
