Hybrid hill-type and reflex neuronal system muscle model improves isometric EMG-driven force estimation for low contraction levels

摘要: Non-invasive muscle force estimation by EMG signals can be obtained using a priori information provided mathematical models of dynamics. In this study, EMG-driven isometric estimates performed three model formulations are compared to verify contribution stretch reflex estimation: (1) Winters’ Hill-type with spindles and Golgi tendon organ system; (2) without reflex; (3) Zajac adapted Menegaldo (ZM model). Submaximal plantar flexion torque predictions, estimated the mentioned above, were dynamometer measurements. Surface was collected from gastrocnemius medialis, lateralis, soleus, tibialis anterior muscles 12 volunteers synchronized The experimental protocol consisted sustained contraction intensities 20 60% individual maximum voluntary (MVC) assisted real-time visual feedback. results show improvement prediction accuracy for at 20% MVC, which not observed MVC.