Learned state prior improves realism in human walking and running simulations
Closing the realism gap in physics-based gait simulations with a learned state prior
Computational Engineering, Finance, and Science
Summary
Simulating how people walk and run using physics can help study different movement scenarios, but these simulations often don’t look very real. The authors trained a method on many real walking and running examples so the simulations produce more natural movements that closely match real experiments. Their improved approach even handles tricky movements like curved running using only minimal data. This makes it easier to test how changes like muscle weakness or shoe types affect walking or running in a more realistic way.
What this means in practice
- •For biomechanical engineers: Create more accurate physics-based simulations of human walking and running to evaluate interventions or assistive devices.
- •For sports technology developers: Design simulation tools that predict how footwear or muscle conditions change athlete gait for improved product testing.
Authors
Markus Gambietz, Zhihao Zhao, Theodoros Balougias, Xiang Wang, Anne D. Koelewijn
Abstract
Predictive simulation of human movement is a promising tool for studying ``what-if'' scenarios in human movement and its underlying motor control, yet its realism is often limited. To address this gap, we incorporate a learned state prior that is trained on a large-scale dataset of human gait kinematics and external forces into predictive simulations. Resulting gait simulations yield kinematics and kinetics across diverse walking and running speeds that better match experimental data than current physics-based simulations, achieving accuracy comparable to data-driven models that reproduce learned data. Furthermore, our method enables robust hypothesis testing by demonstrating how varying optimality assumptions, muscle weakness, and footwear choices influence predicted gait. We also show that this prior generalizes well beyond its training data, successfully reconstructing full-body kinematics for curved running and cutting maneuvers from sparse marker sets. Ultimately, these results suggest that state priors should be broadly integrated into predictive simulations.