Effects of Cost Structure in Optimal Control on Biological Arm Movement: A Simulation Study

Abstract

We have to choose muscle activation pairs of agonist and antagonist muscles from a variety of combinations to achieve a movement. Even though there is a redundancy problem, we could immediately solve the problem and generate movements with a characteristic muscle activation pattern that the muscle pairs burst alternatively as the biphasic or triphasic shape. In this paper, in order to investigate requirements that derive the muscle activation pattern, we carried out numerical simulations of arm movement using a musculoskeletal arm model and an approximately optimal feedback control law with changing the cost structure. As a result, the muscle activation pattern could be reproduced by the simulation with a cost form composed by four terms, i.e., position, velocity, force and energy consumption. Thus, the muscle activations may correspond to cost terms. Furthermore, we suggest that the brain also regulate the force as well as the spatial accuracy and efficiency in the absence of any force interaction.