Differentiable Filtering for Inertial Odometry
Differentiable Filtering for Inertial Odometry
This paper introduces Differentiable Filtering for Inertial Odometry (DFIO), a learning framework that embeds error-state Kalman filter update stages directly into a computational graph.
The authors present DFIO, a framework embedding error-state Kalman filter updates into a computational graph to jointly optimize displacement and covariance heads via differentiable measurement updates, validated on public pedestrian datasets with calibrated and uncalibrated inertial measurements.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- paper: not_reported
- code: Reported pending Open source S5
- weights: not_reported
- dataset: not_reported
- code: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
lts on a pedestrian dataset demonstrate that DFIO reduces mean absolute and one-second relative translation errors by 13.2% and 39.6% relative to the TLIO baseline, and by 7.7% and 17.4% relative to its pre-trained counterpart. Furthermore, evaluations on uncalibrated smartphone inertial data indicate that the proposed method effectively mitigates translational drift and maintains trajectory alignment under highly no
Open source S4
isy sensor conditions. Source code will be made publicly available.
Open source S5
Source：arXiv Robotics — research abstracts · arxiv.org