Embodiment Shapes Rolling Behavior in a Multimodal Infant Model

2026-06-15

AI Summary

This study investigates how early infant rolling behavior is shaped by the interplay between body morphology and the sensorimotor system. Leveraging the embodied, multimodal virtual infant model MIMo—augmented with proprioceptive and vestibular sensing—the authors employ reinforcement learning to simulate the transition from supine to prone posture. For the first time, this work demonstrates through embodied computational modeling how dynamic changes in body morphology influence motor development, successfully replicating the age-related improvements in rolling efficiency and coordination observed in real infants. The emergent behavioral patterns exhibit strong alignment with empirical developmental data, underscoring the critical role of morphological maturation in shaping early motor skills.

Abstract

Rolling over is one of the earliest milestones in infant motor development, reflecting the emergence of coordinated, whole-body sensorimotor control. Here, we conduct a computational study of infant rolling using MIMo, a virtual infant embodiment equipped with proprioception and vestibular sensation. MIMo learns supine-to-prone rolls with reinforcement learning. Interestingly, the learned behaviors capture developmental trends and coordination patterns consistent with those reported in real infants, including improved performance and faster execution with age. Our results explain how infant capabilities and constraints can give rise to realistic behaviors in artificial agents, with a particular emphasis on how motor development is shaped by the changing body morphology. This work highlights the role of embodied computational models as a powerful tool for studying sensorimotor development.

Problem

• embodiment
• infant motor development
• rolling behavior
• sensorimotor control
• body morphology

Innovation

• embodied AI
• reinforcement learning
• infant motor development
• sensorimotor control
• computational modeling