Group Inertial Poser: Multi-Person Pose and Global Translation from Sparse Inertial Sensors and Ultra-Wideband Ranging

TL;DR

This work introduces Group Inertial Poser (GIP), a method for robust multi-person 3D pose and global translation estimation using sparse IMUs augmented by inter-sensor distances from UWB. It combines structured state-space models for per-user pose estimation with a two-step optimization (initial relative-position alignment and trajectory refinement) to enforce inter-person distance constraints in a shared world frame, without requiring calibrated starts. The authors present GIP-DB, the first IMU+UWB dataset for two-person interactions, and demonstrate superior accuracy and robustness over synthetic and real-world data, including multi-person scenarios up to four participants. Collectively, GIP advances wearable-motion capture by enabling reliable, interaction-aware tracking in unconstrained environments and paves the way for practical, real-world applications of IMU+UWB fusion in crowd dynamics and social interaction analysis.

Abstract

Tracking human full-body motion using sparse wearable inertial measurement units (IMUs) overcomes the limitations of occlusion and instrumentation of the environment inherent in vision-based approaches. However, purely IMU-based tracking compromises translation estimates and accurate relative positioning between individuals, as inertial cues are inherently self-referential and provide no direct spatial reference for others. In this paper, we present a novel approach for robustly estimating body poses and global translation for multiple individuals by leveraging the distances between sparse wearable sensors - both on each individual and across multiple individuals. Our method Group Inertial Poser estimates these absolute distances between pairs of sensors from ultra-wideband ranging (UWB) and fuses them with inertial observations as input into structured state-space models to integrate temporal motion patterns for precise 3D pose estimation. Our novel two-step optimization further leverages the estimated distances for accurately tracking people's global trajectories through the world. We also introduce GIP-DB, the first IMU+UWB dataset for two-person tracking, which comprises 200 minutes of motion recordings from 14 participants. In our evaluation, Group Inertial Poser outperforms previous state-of-the-art methods in accuracy and robustness across synthetic and real-world data, showing the promise of IMU+UWB-based multi-human motion capture in the wild. Code, models, dataset: https://github.com/eth-siplab/GroupInertialPoser

Figures (6)

• Figure 1: Our method Group Inertial Poser estimates 3D full-body poses and global translation for multiple humans using inertial measurements from a sparse set of wearable sensors, augmented by the distances between the sensors via ultra-wideband ranging. Our approach overcomes the challenge of drift in previous inertial pose estimators to track translation, as we leverage information from body motions across multiple people. Our IMU+UWB method stabilizes and improves individual pose estimates, relative translation estimates between people, and global translation estimates, thereby preserving meaningful interaction dynamics.
• Figure 2: Overview of Group Inertial Poser (GIP). Our pipeline consists of three key steps. It begins with individual pose estimation using an SSM-based model $\mathbf{\mathcal{E}_\theta}$ to generate a full-body SMPL pose and translations. The optimization steps then refine these estimates by minimizing the discrepancy between predicted and actual between-sensor distances. First, GIP optimizes the initial relative positions ($T_0^{12}$); second, it fine-tunes the translations for both users ($T^1$, $T^2$).
• Figure 3: Our SSM-based model takes the global acceleration, rotation, and between-sensor distances to estimate the pose and translation via SMPL.
• Figure 4: (a) Our dataset includes pairs of participants equipped with (b) various motion capture sensors. These sensors capture acceleration, orientation, and distance data, along with ground-truth SMPL pose parameters and translations for each participant.
• Figure 5: Comparison of translation error on InterHuman.