ESTIMATING COMPLETE GROUND REACTION FORCES AND ANKLE JOINT TORQUES FROM PRESSURE INSOLE DATA IN WALKING AND RUNNING

Abstract

This study described a method to estimate complete ground reaction forces and ankle joint torques from only pressure insole sensor data. One subject performed 56 valid steps during heel-toe walking and running in canvas shoes at different speeds (0.85-4.00m/s). Motion data, ground reaction force data and plantar pressure data were simultaneously recorded. Six parameters of ground reaction forces and ankle joint torques were obtained, and were trimmed from touchdown to takeoff. Pressure values at each of the 99 sensors were used to predict these six parameters by linear stepwise regression. Prediction models with 80% or more explained variance were constructed. Estimated values were obtained from these models, and the prediction accuracy was reported. Comparison to a previously developed method was done.