AUTHOR=Copaci Dorin , Moreno Luis , Blanco Dolores 

TITLE=Two-Stage Shape Memory Alloy Identification Based on the Hammerstein–Wiener Model

JOURNAL=Frontiers in Robotics and AI

VOLUME=Volume 6 - 2019

YEAR=2019

URL=https://www.frontiersin.org/journals/robotics-and-ai/articles/10.3389/frobt.2019.00083

DOI=10.3389/frobt.2019.00083

ISSN=2296-9144

ABSTRACT=Thanks to his characteristics such as high force and light weight, a good biocompatibility, noiseless operation and simplicity, relatively low-cost compared with other conventional actuators; the actuators based on shape memory alloy are currently one of the interest research topics, introducing them in application such robotics, medicine, automation etc.  For a good actuator integration on these types of applications, a proper control is needed, which seems to be no trivial task due his hysteresis, dilatory response and nonlinear behaviour. This work presents a new form of modelling this type of actuator based on Hammerstein - Wiener models, identified on two stages of his operation, when the activation temperature for the actuator is obtained by the Joule effect: electrically energy was transformed in thermal energy, and the second stage when the thermal energy is transformed in mechanical work. To fulfil this objective experimental data such as the input signal (pulse-width modulation), temperature signal and position signal, from the two stage was obtained for a specific shape memory alloy wire, and specific environmental conditions. This data was used in the modelling process. The final model consists on a combination of the models from the two stages, represent the behaviour of the shape memory alloy actuator where the input signal is the pulse-width modulation signal and the output signal are the position of the actuator. The results indicate that the identified model has a very similar response to the behaviour of the real actuator. This model offers the possibility to be used to tune different control algorithms, simulate the entry system before to manufacture them and test on real devices.