Biomimetic Control of Mechanical Systems Equipped with Musculotendon Actuators

摘要

This paper addresses the problem of modelling, control, and simulation of a mechanical system actuated by an agonist-antagonist musculotendon subsystem. Contraction dynamics is given by case I of Zajac's model. Saturated semi positive proportional-derivative-type controllers with switching as neural excitation inputs are proposed. Stability theory of switched system and SOSTOOLS, which is a sum of squares optimization toolbox of Matlab, are used to determine the stability of the obtained closed-loop system. To corroborate the obtained theoretical results numerical simulations are carried out. As additional contribution, the discussed ideas are applied to the biomimetic control of a DC motor, i.e., the position control is addressed assuming the presence of musculotendon actuators. Real-experiments corroborate the expected results.