A hybrid muscle-in-the-loop robot system for studying the neuromechanical properties of movement

Abstract

Body stability and, to a greater extent, locomotion in animals involve complex nonlinear interactions between the body and its natural environment; therefore, in order to fully understand physiological movement strategies, we must first quantify the mechanical properties of muscles as they interact with the environment. The purpose of this study was to develop a closed-loop neuromechanical system that applies real-time control to couple an isolated muscle to a physical environment using a robotic device. Our model system consists of a modified four-bar linkage that facilitates variations in stance width and center of mass (CoM) position; a closed-loop control system that virtually embeds a frog muscle into the robot; and a custom compliant multielectrode array (MEA) that facilitates muscle stimulation. Here we demonstrate a very simplistic balance control mechanism using muscle co-contraction in order to validate our hybrid robot system. As was expected, the kinematics of the hybrid robot when subjected to lateral perturbation were representative of increasing stiffness of the hip joints as the amplitude of stimulation was increased. The system was also able to capture the nonlinear relationship between muscle force and length as well as between force and velocity, which can play an important role in balance. Our current system is limited, however, to embedding a single muscle at a time, which we found could produce undesirable asymmetry in the CoM position during perturbation. Although there is room for improvement, this work provides a useful proof-of-concept that lays a foundation for the development of future closed-loop neuromechanical systems.