Walking Assist Device Control without Delay Based on Plantar Forces Prediction by IMU Sensors

Abstract

To enable high-fidelity control of a walking assist device, it is effective to predict the future human motion from wearable sensors. Since changes in plantar forces are related to the walking cycle, we focus on planter forces to control. In this research, we proposed a method for predicting future plantar forces from the data measured by IMU sensors. We measure the relationship between plantar forces and IMU data during walking. Long Short-Term Memory (LSTM) is used for the prediction. In the experiments, the performance of the proposed method is confirmed for different predicted time. We also test to control a walking assist device based on predicted plantar forces.