AHRS-Assisted Dynamic Response Compensation of VGRF for Capacitive Force-Sensing Insoles

Due to the high accuracy and repeatability, capacitive force sensor array-based smart insoles enable clinical-grade mobile motion assessments for healthcare professionals, trainers, and researchers. Nonetheless, the capacitive sensors exhibit inferior performance in dynamic response, a critical indicator for motion assessment. Conversely, inertial sensing modules (ISMs) possess exceptional dynamic response capability. Accordingly, this study developed a pair of smart insoles, each containing a capacitive force sensor array. The dynamic response of the left insole was compensated by the accelerations from an ISM, i.e., attitude and heading reference system (AHRS), attached to the subject’s left shank. A radial basis function neural network (RBFNN) incorporating prior knowledge, i.e., the principle of a capacitive force sensor, was adopted to calculate the vertical ground reaction force (vGRF) using force data obtained from the force sensor array. According to Newton’s second law of motion, the vGRF is also estimated by multiplying the vertical linear acceleration and half of the test subject’s body weight. A Kalman filter (KF) was implemented as the fusion algorithm. The proposed method was practically tested by comparing it to a 3-D high-precision force plate from Kistler Group, where the test subjects performed strenuous movements, including rapid jumping. The root-mean-square error (RMSE) of the transient impact’s vGRF was 358.9 ± 40.7 N, indicating a 49.1% increase in accuracy compared to that of the capacitive sensor insole. Experimental results suggest that the proposed method improves the measurement performance of capacitive sensing insoles, thereby significantly broadening their applicability.