Inertial Tracking Framework for Shared Bikes in GNSS-Blocked Areas

A research paper on arXiv proposes an inertial tracking framework for large-scale shared bikes in environments where Global Navigation Satellite Systems (GNSS) are unavailable, such as urban canyons. The framework integrates bicycle mechanical constraints with a mixture-of-experts model to address challenges like cumulative drifts and poor robustness from low-cost inertial sensors. It uses multiple expert modules to capture shared representations, weighted by a gating mechanism, improving multi-task learning and enabling uncertainty-aware trajectory estimation. The method avoids reliance on visual or LiDAR sensors, which are unsuitable for large-scale deployment. The paper is authored by researchers and published on arXiv with ID 2605.07412.