\textbf{PLATO}: \emph{Preintegration Learning from Accurate Trajectory Observations} for Neural Inertial Odometry

\textbf{PLATO}: \emph{Preintegration Learning from Accurate Trajectory Observations} for Neural Inertial Odometry

|Authors: Haoying Li, Qihang Liu, Yifan Peng, Keyan Miao, Junfeng Wu

|Venue: arXiv preprint | Year: 2026

|arXiv: 2609.06450v1

Abstract

Neural inertial odometry has demonstrated strong potential for motion estimation in challenging environments, yet inertial-only preintegration remains sensitive to IMU bias and uncertainty. To this end, this paper introduces \textbf{PLATO}:\emph{Preintegration Learning from Accurate Trajectory Observations}, a likelihood-based framework that leverages accurate trajectory observations to jointly learn IMU bias dynamics modeled by a neural ordinary differential equation(NODE) and gyroscope and accelerometer noise covariances. Optimization exploits the sparse structure of the negative log-likelihood, with IMU noise-parameter gradients computed by forward differentiation. A tailored double-adjoint scheme couples a discrete invariant-error adjoint with a continuous-time adjoint for the bias NODE, enabling memory-efficient likelihood optimization over the nested bias-dynamics and IMU-preintegration rollouts. Validation on EuRoC shows improved performance, and underwater robot experiments demonstrate applicability under intermittent lighting failures and visual degradation.

Key Contributions

• Neural inertial odometry has demonstrated strong potential for motion estimation in challenging e…
• To this end, this paper introduces \textbf{PLATO}:~\emph{Preintegration Learning from Accurate Tr…
• Optimization exploits the sparse structure of the negative log-likelihood, with IMU noise-paramet…
• A tailored double-adjoint scheme couples a discrete invariant-error adjoint with a continuous-tim…
• Validation on EuRoC shows improved performance, and underwater robot experiments demonstrate appl…

Topics

• vision

Code & Data

No code repository linked in paper metadata.

BibTeX

@article{Li2026_260906450v1,
  title     = {\textbf{PLATO}: \emph{Preintegration Learning from Accurate Trajectory Observations} for Neural Inertial Odometry},
  author    = {Haoying Li and Qihang Liu and Yifan Peng and Keyan Miao and Junfeng Wu},
  year      = {2026},
  eprint    = {2609.06450v1},
  archivePrefix = {arXiv},
  primaryClass  = {cs.RO},
  url       = {https://arxiv.org/abs/2609.06450v1}
}