Humans coactivate agonist-antagonist muscles to modulate the limb impedance (stiffness, damping, inertia) in a time- and task-dependent manner, independently from the kinematics of the limb. Estimation of the motor intent in terms of joint kinematics and impedance would therefore be relevant when substituting missing limbs with artificial ones.
We present AIC-UP, a novel framework for Adaptive Impedance Control of Upper-Limb Prosthesis that uses muscle-tendon models, driven by surface electromyographic signals from agonist-antagonist muscle groups, to enable voluntary control of the kinematics, stiffness and damping of a Degree of Freedom of a simulated robot. AIC-UP does not require measurement of joint torque or stiffness to train the models and it is therefore suitable for application in upper-limb prosthesis. This was achieved through i) reparameterization of muscle-tendon unit models, ii) the inclusion of structural assumptions to constrain the parameter space of the models, and iii) the formulation of an optimization framework for the training muscle-tendon models that includes a position-based impedance controller.
AIC-UP was evaluated with eight able-bodied subjects performing target-reaching tasks in simulation through wrist flexion-extension. The control performance was tested in free space and in the presence of unexpected perturbations.
We showed that AIC-UP outperformed a neural network that regresses the desired kinematics from EMG signals and allowed the able-bodied participants to exploit joint stiffness and damping adaptation to ensure stable interaction with the environment and successful task execution. While the methods were tested with a single amputee, the obtained results were coherent with those of the able-bodied participants.