Virtual Track: A Vision-based Integrity Enhancement

The upcoming transport revolution is dramatically increasing the navigation integrity requirements. However, traditional techniques for improving the integrity of the position information are usually based on the use of a high number of satellite signals. We propose to add an external integrity source based on the use of on-board visual sensors. Shift and rotation of the vehicle are estimated from the vanishing point of the road markings, and exploited to define a virtual track for its motion. The GNSS-only solution is then compared to the one constrained to the virtual track, and an integrity check is performed based on a threshold. The performed experiments prove the possibility of exploiting the visual system to reliably compute the vehicle's dynamics from the acquired images, and show the impact of the virtual track constraint on the integrity check.