Geometric interface improves inertial odometry under sensor rotations
GINIO: A Geometric SO(3)-Equivariant Interface for Neural Inertial Odometry
RoboticsMachine Learning
Summary
Inertial odometry uses sensors to track movement, but measurements can get confused if the sensor is mounted differently. The authors propose GINIO, a system that makes the measurements behave correctly no matter how the sensor is rotated. Their method predicts motion and uncertainty in a way that always matches real-world rotations, making tracking more accurate and reliable even when the sensor is physically moved or remounted. Tests show substantial improvement in positioning accuracy and stability without retraining.
What this means in practice
- •For robotics engineers: Improve robot navigation accuracy by using an inertial odometry interface robust to IMU sensor rotations and remounting.
- •For drone developers: Enhance aerial vehicle position tracking with a neural odometry system that maintains accuracy despite sensor orientation changes.
Authors
Chankyo Kim, Minghan Zhu, Tzu-Yuan Lin, Avantika Rattan, Maani Ghaffari
Abstract
Neural inertial odometry increasingly uses networks as learned measurements inside filtering pipelines. Such measurements should transform consistently under arbitrary IMU mounting conventions: their mean must transform as a vector, and their covariance must transform congruently as a second-order tensor. We present GINIO, a geometric SO(3)-equivariant interface for neural inertial odometry under arbitrary rotations of the IMU measurement frame. Given calibrated IMU windows, our framework predicts a motion measurement and uncertainty obeying these tensorial laws. To support efficient sensor-frame learning, we introduce Last-Frame Alignment (LFA), a deterministic preprocessing step that is provably equivalent to world-frame training for SO(3)-equivariant predictors. The connected estimator tracks sensor-local states such as IMU bias, separating nuisance estimation from the geometric law enforced by the learned measurement. We instantiate the same interface in filter-connected NIO, AirIO-style recurrent aerial prediction, EqNIO-style full-SO(3) canonicalization, and ResNet-style temporal backbones. On TLIO, GINIO achieves 2.018 m ID/SO(3) ATE while EqNIO degrades to 76.389 m, using 11.6x fewer FLOPs. On NanoBench, our AirIO-style instantiation improves ATE from 5.579 m to 1.430 m without external attitude input, and our ResNet-style instantiation reaches 0.581 m ATE versus 0.645 m for ResNet1D. On Fetch, GINIO empirically reduces unseen physical-remount ATE from 8.15 m to 0.50 m without retraining, demonstrating robustness beyond the exact coordinate-frame guarantee. For uncertainty, spectral covariance reduces covariance-equivariance error by over three orders of magnitude compared with a diagonal head.