A Robotic Model of the Human Neuro-Musculo-Skeletal System

A pair of muscles powering the human joint in an antagonistic configuration exemplifies the main difference between standard industrial robots and biological motor systems. Since muscles have a natural stiffness that varies with the muscle activation level, the central nervous system can generate stable equilibrium postures, towards which the arm is attracted, by properly regulating the activation levels of antagonistic muscles [1] The elastic properties of muscles contribute to the finite stiffness/compliance properties of the limb, to the stability of the neuro-musculo-skeletal system in the face of significant feedback delays and even allow for the generation of target movements in absence of sensory feedback, by shifting the equilibrium point [2]. Control theories based on the presence of the EP in biological motor systems [3] suggest that movements are programmed as a shift of equilibrium positions rather than through an explicit computation of forces. Thus, there is no need to solve the “inverse dynamics problem” for calculating the torque required to move the arm on the desired trajectory. The implementation of a given neuroscientific hypothesis on a real mechanical system could provide a tool under the full control of the experimenter, reproducing the main functional features of the human arm and being able to interact with the same physical environment of the human. To this end we developed the NEURARM platform, a bio-mimetic planar robotic arm reproducing key features of the human arm as identified at the level of joints and muscles. In particular