Related Experiment Video
Updated: Oct 28, 2026
Simulation of Human-induced Vibrations Based on the Characterized In-field Pedestrian Behavior
Published on: April 13, 2016
A novel force scaling method for ground reaction force estimation from pressure insoles to enable long-term remote
Karl E Zelik1,2,3, Katherine M Rodzak1, Meredith K Owen4
1Department of Mechanical Engineering, Vanderbilt University, Nashville, Tennessee, United States of America.
Background:
Pressure-sensing insoles offer a portable way to quantify human movement biomechanics beyond the lab, enabling real-world monitoring. A common objective with these insoles is to estimate ground reaction forces (GRFs). However, force estimates from insoles typically exhibit a magnitude bias-underestimating or overestimating GRFs-that can vary across individuals and contexts. Conventional force scaling approaches can correct for this bias, but they rely on ground truth force plate data for calibration. This is often impractical for remote or longitudinal data collection, particularly for individuals with lower-limb injuries or disabilities.
Objective:
The purpose of this study was to introduce and evaluate a new force scaling method for pressure insoles to enable more accurate estimation of vertical GRFs without laboratory calibration or access to ground truth force plate data, and which could be applied to individuals with lower-limb injuries or disabilities.
Methods:
The new method combines epoch-based and physics-based approaches to perform periodic and dynamic force scaling based on a user's body weight, in a way that is well-suited for use in real-world patient monitoring, and which does not require an individual to weight-bear on a single limb. This force scaling method was evaluated across different participants and activities representative of daily living and rehabilitation exercises.
Results:
Compared to nominal (unscaled) insole force estimates, the new scaling method reduced GRF errors by an average of 54-72%, achieving average errors on the order of 0.11-0.16 body weights. The new method worked reliably for different individuals, both limbs, and a variety of tasks and exercises.
Conclusions:
This method provides a practical, generalizable, and effective approach for improving vertical GRF estimation from pressure insoles during remote, longitudinal monitoring without in-lab calibration or force plates. This could improve GRF assessment outside clinical or laboratory settings and enable clinicians, researchers, and users to more accurately monitor musculoskeletal loading and other biomechanical metrics during rehabilitation, recovery, and physical activities in real-world environments.
