Organizations: I3S, CNRS, Université Côte d’Azur 06900 Sophia Antipolis, France · Département d’informatique et d’ingénierie, Université du Québec en Outaouis Gatineau, QC J8X 3X7, Canada · Department of Electrical Engineering, Lakehead University Gatineau, QC J8X 3X7, Canada · Institut Universitaire de France (IUF)
This paper presents a geometric observer for monocular visual-inertial odometry driven by individual optical-flow constraints. For each tracked point, the bearing and its rotationally compensated derivative give a depth-free homogeneous constraint on the body-frame velocity. These constraints enter the innovation of an observer on the gravity-augmented group SE3(3) through a Riccati gain, without estimating depths or reconstructing the direction of motion. The auxiliary gravity state makes the six-dimensional body-frame velocity-gravity error dynamics independent of the attitude correction, and their uniform observability is shown to be equivalent to persistent excitation of a matrix formed from the inertial bearings and their derivatives. Under this condition, the body-frame velocity and gravity errors converge globally and uniformly exponentially, and a gravity-based attitude correction renders the combined velocity, gravity, and reduced-attitude error almost globally asymptotically stable and locally uniformly exponentially stable. Simulations illustrate the excitation condition, recovery from initial errors, and sensitivity to perturbed measurements.
Figures & tables
Figure 1: Optical-flow-based VIO. The onboard camera observes static landmarks, whose apparent image motion is estimated by optical-flow, while the IMU provides the vehicle’s inertial measurements.
Figure 2: Depth-free constraint from one feature with si=0 , viewed in the plane νi⊥ . The circle is a section of the unit bearing sphere; the dashed tangent contains both si and πbivB , which point in opposite directions.
Figure 3: Observer construction on SE3(3) . The measured flow coefficients and the estimated inverse group column form the visual residual. The residual, the Riccati gain, and the estimated gravity column determine the Lie-algebra innovation, which corrects the group state.
Figure 4: Ideal measurements, n=8 : velocity-error norm and minimum eigenvalue of Nδ(t):=δ−1∫t−δtN(s)ds , with δ=4πs (complete windows only). Velocity errors are floored at 10−8m/s for display. Straight translation violates the excitation condition despite the changing attitude.
Figure 5: Recovery on the figure-eight, n=8 : 100 randomized initial states per condition, noise-free and at nominal noise. Curves are pointwise medians, bands empirical 5th–95th percentiles; values below 10−8 are floored. Reduced attitude excludes the unobservable yaw.
Figure 6: Matched-initialization odometry on the figure-eight, n=8 , at nominal and fourfold noise (100 trials each) and without noise. Position error is in the common initial frame, unaligned.
Computer Science, Maharishi International University, Iowa, USA · Electrical Engineering and Computing, University of Zagreb, Zagreb, Croatia · Computer Science, Iowa State University, Iowa, USA +1
Shanghai Institute of Technical Physics, Chinese Academy of Sciences, Shanghai 200083, China · University of Chinese Academy of Sciences, Beijing 100049, China · GravityXR, Ningbo 315040, China +2