Reaching Motion Characterization Across Childhood via Augmented Reality Games

TL;DR

Reaching Motion Characterization Across Childhood via Augmented Reality Games investigates how upper-limb reaching strategies develop in children using an AR game dataset. The authors show that short 13-second clips suffice to differentiate developmental stages and reveal systematic changes: directness of reach increases with age, peak velocity declines, and anticipatory reaching improves, reducing overshoot. They introduce directness, velocity, and progress-rate metrics, apply 3D reconstruction to validate findings, and demonstrate a temporal convolutional network can predict a child’s age from arm motion with RMSE of $2.77$ years. The work suggests AR/VR game-based motion capture can serve as a scalable diagnostic tool and contribute to realistic animation of child characters.

Abstract

While performance in coordinated motor tasks has been shown to improve in children as they age, the characterization of children's movement strategies has been underexplored. In this work, we use upper-body motion data collected from an augmented reality reaching game, and show that short (13 second) sections of motion are are sufficient to reveal arm motion differences across child development. To explore what drives this trend, we characterize the movement patterns across different age groups by analyzing (1) directness of path, (2) maximum speed, and (3) progress towards the reaching target. We find that although maximum arm velocity decreases with age (p~=~0.02), their paths to goal are more direct (p~=~0.03), allowing for faster time to goal overall. We also find that older children exhibit more anticipatory reaching behavior, enabling more accurate goal-reaching (i.e. no overshooting) compared to younger children. The resulting analysis has potential to improve the realism of child-like digital characters and advance our understanding of motor skill development.

Figures (5)

• Figure 1: AR game data processing pipeline. (left) Reaching mini-game screenshot of participant bilaterally reaching towards virtual targets displayed on a screen. (center) Illustrated upper body skeleton overlaid on video frame. (right) Extracted skeleton.
• Figure 2: Sample reach trajectories of one arm during the same bilateral reach (same target positions) from each age group. The sampled trajectories vary in strategy, demonstrating differences in speed, directness, and time to reach-completion. The older child's quick and direct path is in stark contrast with the younger children.
• Figure 3: Confusion matrix of age regression binned across four age groups. Youngest and oldest age groups are most clearly differentiable.
• Figure 4: (a) Participants display more direct motion as they age (shorter paths taken to goals), (b) but move slower, which indicates more controlled motion.
• Figure 5: Splines showing normalized reaching progress curves averaged across three age bins.