Low-Complexity Output Feedback Control With Prescribed Performance for Bioinspired Cable-Driven Actuator

This article addresses the modeling and control of bioinspired cable-driven actuator that can mimic the three activation modes of human muscles. First, the dynamic model of the actuator is established by utilizing the coordinate transformation and kinetic energy theorem for controller and observer design. Then, a low-complexity output feedback controller with prescribed performance is proposed, so that the driving disc can track the rotation of the cable pulley without overshoot during the operation of the actuator. Specifically, the collision caused by overshoot during the tracking process must be avoided, as it can cause a sudden change of the cable tension, impact noise, and even mechanical parts damage. The presented low-complexity control strategy guarantees that the tracking error does not violate the predefined performance constraint and can converge to a predetermined residual set in finite time by introducing a novel error transformation function and a performance function. In addition, to avoid using angular velocity sensor, a finite-time observer with interval type-2 fuzzy system and adaptive technique is developed to estimate the angular speed of the driving disc. Finally, simulation studies and prototype experiments are performed to validate the effectiveness of the proposed control scheme on the bioinspired actuator.