MosaicIMU: Composing Carrier Experts for Generalizable Neural Inertial Odometry
2026-06-08 • Robotics
Robotics
AI summaryⓘ
The authors developed MosaicIMU, a new method for tracking movement using sensors inside devices when outside signals are unreliable. It uses a special system that learns patterns from different types of carriers (like phones or robots) and adapts to new devices by adding small updates without retraining everything. This approach mixes expert models to better predict velocity and uncertainty and improves accuracy by combining these predictions with a smart filter. Their tests showed it works better than other learning-based methods and can easily adjust to new platforms while being efficient for real-time use.
inertial odometryMixture-of-Experts (MoE)carrier-conditioned modelingprototype-based routingExtended Kalman Filter (EKF)local velocity estimationdomain adaptationincremental learningintegration driftneural sensor fusion
Authors
Junye Zou, Huiyi Yan, Xinning Xu, Xiaolei Li, Pengkun Zhou, Jinhui Zhang, Ziyang Meng
Abstract
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.