NANO-L is a natural-gradient Gaussian approximation filter on Lie groups that avoids linearization by optimizing multiplicative increments via the exponential map, yielding exact covariance updates for invariant models and 40% lower error on hardware.
Title resolution pending
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
cs.RO 3years
2026 3verdicts
UNVERDICTED 3representative citing papers
Tangent-linearized Gaussian inference on manifolds has explicit non-asymptotic W2 stability bounds that predict a calibration transition near sqrt of the operator norm of covariance over reach approximately equal to 1/6.
The NANO filter uses natural gradient descent to iteratively refine Gaussian state estimates while preserving covariance positive definiteness and exactly recovering the Kalman update in the linear-Gaussian case.
citing papers explorer
-
Natural Gradient Gaussian Approximation Filter on Lie Groups for Robot State Estimation
NANO-L is a natural-gradient Gaussian approximation filter on Lie groups that avoids linearization by optimizing multiplicative increments via the exponential map, yielding exact covariance updates for invariant models and 40% lower error on hardware.
-
Distributional Stability of Tangent-Linearized Gaussian Inference on Smooth Manifolds
Tangent-linearized Gaussian inference on manifolds has explicit non-asymptotic W2 stability bounds that predict a calibration transition near sqrt of the operator norm of covariance over reach approximately equal to 1/6.
-
Natural Gradient Bayesian Filtering: Geometry-Aware Filter for Dynamical Systems
The NANO filter uses natural gradient descent to iteratively refine Gaussian state estimates while preserving covariance positive definiteness and exactly recovering the Kalman update in the linear-Gaussian case.