#SeminarIRSJD · Simulation of Assisted Motion with Musculoskeletal Models
- Dr. Albert Peiret Giménez · Investigador del grup Tecnologies mèdiques: Biomecànica · Institut de Recerca Sant Joan de Déu.
Seminari exclusiu per personal de la institució. Si desitges obtenir l'enllaç de connexió contacta amb info(ELIMINAR)@irsjd.org
Resum
Robotic assistive devices such as exoskeletons can help patients to walk again after a spinal cord injury. The control of such devices and their adaptation to the patient's motion for an optimal assistance are still unsolved problems. Approaches based on optimal control techniques with a patient-specific musculoskeletal model can help to make predictions on the patient's motion and optimise the rehabilitation process. Furthermore, other approaches based on reinforcement learning cap obtain robust solutions for not only the device controller but also the neural system using artificial neural networks. The cornerstone of these techniques is the minimisation a cost function, which can represent muscle metabolic energy as well as other motion task goals. Modelling muscle metabolic energy is especially useful for other forms of assistance based on functional electrical stimulation, which induces muscle contraction to generate functional motion.