UMotion: Uncertainty-driven Human Motion Estimation from Inertial and Ultra-wideband Units

TL;DR

UMotion introduces an uncertainty-driven framework that fuses six IMU-UWB sensors for online 3D human shape and pose estimation. The approach jointly learns body shape via an ensemble method using anthropometrics and inter-sensor distances, and predicts pose with a unidirectional LSTM that outputs both pose parameters and associated uncertainty. A tightly coupled Unscented Kalman Filter then refines IMU, UWB, and pose information in real time, propagating uncertainties through SMPL constraints to stabilize distances and poses. Across synthetic and real datasets, UMotion surpasses state-of-the-art IMU-only and distance-augmented methods in pose and mesh accuracy, while maintaining robustness to occlusions and sensor noise, enabling practical, continuous motion capture outside lab environments.

Abstract

Sparse wearable inertial measurement units (IMUs) have gained popularity for estimating 3D human motion. However, challenges such as pose ambiguity, data drift, and limited adaptability to diverse bodies persist. To address these issues, we propose UMotion, an uncertainty-driven, online fusing-all state estimation framework for 3D human shape and pose estimation, supported by six integrated, body-worn ultra-wideband (UWB) distance sensors with IMUs. UWB sensors measure inter-node distances to infer spatial relationships, aiding in resolving pose ambiguities and body shape variations when combined with anthropometric data. Unfortunately, IMUs are prone to drift, and UWB sensors are affected by body occlusions. Consequently, we develop a tightly coupled Unscented Kalman Filter (UKF) framework that fuses uncertainties from sensor data and estimated human motion based on individual body shape. The UKF iteratively refines IMU and UWB measurements by aligning them with uncertain human motion constraints in real-time, producing optimal estimates for each. Experiments on both synthetic and real-world datasets demonstrate the effectiveness of UMotion in stabilizing sensor data and the improvement over state of the art in pose accuracy.

Figures (12)

• Figure 1: UMotion integrates IMU-UWB data inputs and pose outputs uniformly under uncertainty, constrained by individual body structure. The online state estimation framework iteratively refines sensor data confidence and stabilizes pose estimation, reducing ambiguities and improving robustness.
• Figure 2: Overview of UMotion, consisting of three main modules: the shape estimator, pose estimator, and state estimator. The shape estimator takes anthropometric measurements and inter-distances in a T-pose as input, outputting shape parameters that reconstruct a realistic body and impose strong constraints on the system. The pose estimator receives filtered IMU data and inter-distances from the state estimator to predict poses, which are fed back to refine state estimates. The entire system integrates IMU, UWB, and estimated poses within the context of individual body structure to continuously update and improve motion estimation.
• Figure 3: Visualization of selected inter-sensor distances used in the shape estimator. We place virtual sensors on the body mesh at sensor mounting points and conduct line-of-sight simulation experiments. The plots display the temporal changes in UWB measurements between various sensor pairs over time.
• Figure 4: Cumulative distribution of distance error (left) and acceleration error reduction over time (right) for various fusion settings.
• Figure 5: Joint positional error for different fusion settings.