Answers are grounded in this paper and cite their sources.
Evaluates sensor configurations for estimating ground reaction forces during various movements, indicating efficiency gains.
Ground reaction force (GRF) is an important kinetic variable for analyzing human movement and musculoskeletal function. Although GRF is typically measured using force plates, their use is limited to laboratory environments. To overcome this limitation, recent studies have attempted to estimate GRF using inertial measurement units (IMUs). However, the effects of sensor configuration on estimation performance have not been fully investigated, particularly across different types of movements. In this study, we implemented a recurrent neural network model to estimate three-dimensional GRFs from segmental kinematics obtained from IMUs and systematically evaluated various sensor configurations, including full-body (FB) and reduced configurations (thorax, thighs, shanks, and feet), during gait and non-gait movements. Among the reduced sensor configurations, the combinations of thorax, thighs, and shanks for the five-sensor setup and thorax and shanks for the three-sensor setup exhibited superior performances, particularly in non-gait movements, indicating that shank kinematics are a key contributor to GRF estimation. The comparison of all configurations showed that the FB configuration achieved the highest overall accuracy, whereas the reduced sensor configurations showed only approximately 10–16% performance degradation, demonstrating improved efficiency in terms of sensor usage with five or fewer IMUs.
No takes yet. Share an insight, caveat, or question.
Lee et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: