Reliability-Gated Fusion of Consumer Head and Foot IMUs for Lower-Body 3D Pose
What it is
This paper proposes a method for fusing data from consumer-grade head and foot IMUs to estimate 3D lower-body pose without cameras. It introduces a channel-level reliability gating mechanism that learns to trust or ignore specific sensor channels (like firmware-fused orientation) based on their quality, improving robustness against sensor drift, bias, and dropout.
The problem it solves
Consumer IMUs are unreliable due to firmware bias, mounting variance, and signal drift/dropout, which degrades the accuracy of sparse inertial pose estimation.
What you could use it for
- Camera-free motion capture using earbuds and smart insoles
- Robust pose estimation in environments where sensors may fail or drift
- Augmenting sparse inertial data with reliability-aware fusion
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.