Continuous control of lower limb prosthetic and assistive devices is usually performed by mapping electromyography (EMG) activity to joints angular kinematics. However, the inclusion of ground reaction force information within the regulation scheme would be valuable in order to provide a more physiological walking mechanics, thus enhancing patient comfort and usability in real context scenarios. In this paper, we explored the feasibility of a reliable EMG-based estimation of the vertical ground reaction force (VGRF) component during unconstrained walking. The EMG activity from six lower limb muscles was recorded from five healthy subjects who performed four gait trials, and the VGRF was recorded during the stance phase of gait by dynamometric plates. As a regression model, we relied on a LSTM network. Results showed very promising performances for VGRF estimation, without any drop between training and testing (R2 of about 0.90, and a RMSE of 8.7%). The three thigh muscles failed in providing reliable results (R2 <0.7 and RMSE>14%), whereas shank muscles alone provided no significantly different outcomes with respect to using all the muscles, pointing out the possibility of using a reduced EMG recording setup, thus impacting on actual usability. These findings can be considered as a step toward the inclusion of dynamics information in myoelectric volitional control schemes for enhancing movement reconstruction.
Mapping Lower Limb EMG Activity to Ground Reaction Force in Free Walking Condition / Mengarelli, A.; Scattolini, M.; Mobarak, R.; Verdini, F.; Fioretti, S.; Burattini, L.; Tigrini, A.. - (2024). (Intervento presentato al convegno 2024 IEEE International Symposium on Medical Measurements and Applications, MeMeA 2024 tenutosi a nld nel 2024) [10.1109/MeMeA60663.2024.10596870].
Mapping Lower Limb EMG Activity to Ground Reaction Force in Free Walking Condition
Mengarelli A.;Scattolini M.;Mobarak R.;Verdini F.;Fioretti S.;Burattini L.;Tigrini A.
2024-01-01
Abstract
Continuous control of lower limb prosthetic and assistive devices is usually performed by mapping electromyography (EMG) activity to joints angular kinematics. However, the inclusion of ground reaction force information within the regulation scheme would be valuable in order to provide a more physiological walking mechanics, thus enhancing patient comfort and usability in real context scenarios. In this paper, we explored the feasibility of a reliable EMG-based estimation of the vertical ground reaction force (VGRF) component during unconstrained walking. The EMG activity from six lower limb muscles was recorded from five healthy subjects who performed four gait trials, and the VGRF was recorded during the stance phase of gait by dynamometric plates. As a regression model, we relied on a LSTM network. Results showed very promising performances for VGRF estimation, without any drop between training and testing (R2 of about 0.90, and a RMSE of 8.7%). The three thigh muscles failed in providing reliable results (R2 <0.7 and RMSE>14%), whereas shank muscles alone provided no significantly different outcomes with respect to using all the muscles, pointing out the possibility of using a reduced EMG recording setup, thus impacting on actual usability. These findings can be considered as a step toward the inclusion of dynamics information in myoelectric volitional control schemes for enhancing movement reconstruction.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.