A new approach for hybrid (PID + MRAC) adaptive controller applied to two-axes McKibben muscle manipulator: a mechanism for human-robot collaboration

Autor: Eduardo José Lima, Neemias Silva Monteiro, Vinícius Avelino Sena, Marcelo Henrique Souza Bomfim
Rok vydání: 2021
Předmět:
Zdroj: Industrial Robot: the international journal of robotics research and application. 48:836-845
ISSN: 0143-991X
DOI: 10.1108/ir-12-2020-0267
Popis: Purpose This paper aims to present a new approach, called hybrid model reference adaptive controller or H-MRAC, for the hybrid controller (proportional-integral-derivative [PID + MRAC]) that will be used to control the position of a pneumatic manipulator. Design/methodology/approach It was developed a McKibben muscle using nautical mesh, latex and high-density polyethene connectors and it was constructed an elbow manipulator with two degrees of freedom, driven by these muscles. Then it was presented the H-MRAC control law based on the phenomenological characteristics of the plant, aiming at fast response and low damping. Lyapunov's theory was used as the project methodology, which ensures asymptotic stability for the control system. Findings It was developed a precise control system for a pneumatic manipulator and the results were compared to previous research. Research limitations/implications In collaborative robotics, human and machine occupy the same workspace. This research promotes the development of safer and more complacent mechatronic systems in the event of collisions. Practical implications As a practical implication, the research allows the substitution of electric motors by McKibben muscles in industrial robots with high accuracy. Social implications The pneumatic manipulator will make the human-robot physical interaction safer as it can prevent catastrophic collisions causing victims or equipment breakdown. Originality/value When compared to results in the literature, the present research showed a 37.51% and 36.74% lower global error in position tracking than MRAC and Adaptive proportional-integral-derivative (A-PID), respectively, validating its effectiveness.
Databáze: OpenAIRE