Using Wearable Sensors to Estimate Vertical Ground Reaction Force Based on a Transformer

Abstract

In this paper, we present a new method to estimate ground reaction forces (GRF) from wearable sensors for a variety of real-world situations. We address the drawbacks of using force plates with limited activity range and high cost in previous work. We use a transformer encoder as a feature extractor to extract temporal and spatial features from wearable sensors more efficiently. Using the Mean Absolute Percentage Error (MAPE) as the evaluation criterion, the experimental results show that the average error of the predicted values using the transformer as a feature extractor improved by 32% compared to the RNN architecture and by 25% compared to the LSTM architecture. Finally, we use Gate_MSE to solve the problem of a large peak error in GRF prediction. Meanwhile, this paper explores the effect of the number of wearable sensors or wearable modes on GRF prediction.

Extended Application of Inertial Measurement Units in Biomechanics: From Activity Recognition to Force Estimation

“…Vertical acceleration of relatively large mass limbs has been demonstrated to be highly correlated with the vertical GRF [68]. Therefore, most GRF estimation studies attached IMUs to the pelvis [5,56,57], CoM [16], and the sacrum [9,14]. For the estimation of lower limb joint force/moments, IMUs attached to the shank seem to be able to achieve better estimation accuracy [63].…”

“…Five IMUs worn on the pelvis, left thigh, left ankle, right thigh, and right ankle were used to collect kinematics data. The results showed that the average error of the predicted values was improved by 32% when compared to the RNN architecture and by 25% when compared to the LSTM architecture, while using the transformer as a feature extractor [57].…”

“…Comparatively, in previous studies, the sampling frequency of most IMUs for HAR remained at 25-100 Hz. Of note, the sampling rates of more than 50% of the IMU-based force estimation systems were in the range of 100-250 Hz [9,12,13,16,[56][57][58][59][63][64][65]98], with one exception below 100 Hz [5]. Eight force estimation studies used IMUs with a sampling rate above 1000 Hz [4,10,14,60,62,66].…”

Ground Reaction Forces and Joint Moments Predict Metabolic Cost in Physical Performance: Harnessing the Power of Artificial Neural Networks

“…The significance of ground reaction forces (GRFs) in biomechanical analysis is welldocumented, with methodologies ranging from direct force plate measurements to sophisticated estimations using wearable sensors and inertial motion capture systems [3]. Such advancements have democratized gait and movement analysis, extending the reach of biomechanical assessments beyond the confines of specialized laboratories [4].…”