Observability Analysis of Visual-Inertial Odometry with Online Calibration of Velocity-Control Based Kinematic Motion Models

TL;DR

This paper investigates the observability of visual-inertial odometry systems with online calibration, demonstrating how certain constraints and motion models affect the observability of position, orientation, and parameters.

Contribution

It provides a theoretical analysis of observability in VIO with motion constraints and shows how to achieve observability of all parameters through planar motion.

Findings

Global position and yaw are unobservable without rotation.

Roll and pitch become observable with planar motion constraints.

Motion model parameters are observable.

Abstract

In this paper, we analyze the observability of the visual-inertial odometry (VIO) using stereo cameras with a velocity-control based kinematic motion model. Previous work shows that in general case the global position and yaw are unobservable in VIO system, additionally the roll and pitch become also unobservable if there is no rotation. We prove that by integrating a planar motion constraint roll and pitch become observable. We also show that the parameters of the motion model are observable.