MosaicIMU: Composing Carrier Experts for Generalizable Neural Inertial Odometry
For autonomous systems relying on inertial odometry, this work enables a single model to work across different carriers (e.g., drones, vehicles, pedestrians) without retraining, addressing a key generalization bottleneck.
MosaicIMU introduces a carrier-conditioned Mixture-of-Experts framework for neural inertial odometry that generalizes across heterogeneous platforms, reducing average ATE and RTE-10s by 40% and 34% respectively over learning-based baselines.
Robust inertial odometry is essential for various carriers when external sensing is unreliable. Learning-based methods reduce integration drift by capturing local motion priors, but these methods often remain tied to a particular carrier, limiting generalization across heterogeneous platforms. We present MosaicIMU, a carrier-conditioned Mixture-of-Experts (MoE) pretraining-and-adaptation framework for generalizable neural inertial odometry. MosaicIMU uses a prototype-based router to compose carrier-specific expert features, decodes local velocity and uncertainty constraints, and integrates them with a history-aware EKF. For unseen domain adaptation, it freezes the pretrained base model and learns a new lightweight expert residual branch. For edge-deployment, it further reuses the router to select informative online samples for efficient incremental updates. Experiments show that MosaicIMU consistently outperforms learning-based baselines, reducing average ATE and RTE-10s by 40% and 34%, respectively. These results highlight that MosaicIMU provides a scalable pretraining-to-deployment paradigm for generalizable and adaptive neural inertial odometry.