A Gravity-Referenced Moving Frame for Vehicle Path Following Applications in 3D

Abstract

Moving path frames assigned to spatial curves are commonly used in the development of motion control laws for autonomous vehicles. This work presents the Gravity Normal frame, a novel navigation reference frame developed specifically for autonomous vehicle applications. This moving path frame incorporates the knowledge that many autonomous vehicles operate in a gravitational field, and control strategies must account for this. Given a curve in space that represents a desired trajectory, the proposed strategy generates a navigation frame that is well defined regardless of path curvature, while guaranteeing the normal vector is always normal to gravity and hence constrained to the horizontal plane, regardless of path torsion. Due to these characteristics, the Gravity Normal path frame is ideally suited for vehicles with distinct longitudinal and lateral dynamics since the resulting cross-track errors have a precise physical interpretation. The properties of the navigation frame are derived, and its usefulness is showcased through simulation. Finally, its applicability is demonstrated with flight experiments on a fixed-wing unmanned aerial vehicle.