Hysteresis Modeling for Magnetic Shape Memory Alloy Actuator Based on Dynamic Fuzzy Neural Network

Miaolei Zhoua, Chen Zhanga, Yewei Yua, Shouchun Wangb aDepartment of Control Science and Engineering, Jilin University, Changchun 130022, P.R. China bDepartment of Neurology, The First Bethune Hospital of Jilin University, Changchun 130022, P.R. China

Full Text PDF

Magnetic shape memory alloy is a new class of smart metallic materials, which has excellent characteristics of huge strain and fast response. These characteristics make magnetic shape memory alloy-based actuator a potential alternative to replace traditional actuators in the high-precision positioning applications. However, the magnetic shape memory alloy-based actuator has not found its way into micro positioning field due to the obvious hysteresis behavior. In this paper, we present the prototype of the magnetic shape memory alloy-based actuator, and analyze the complex hysteresis nonlinearity between the input signal and output displacement. Then, dynamic fuzzy neural network is first utilized to construct hysteresis model for the magnetic shape memory alloy-based actuator. Dynamic fuzzy neural network is a fuzzy-logic based neural network system, which has the capability of approximating nonlinear mapping and self-adjustment. Experimental results confirm the effectiveness of the proposed hysteresis model.

DOI:10.12693/APhysPolA.137.660 topics: hysteresis nonlinearity, magnetic shape memory alloy-based actuator, dynamic fuzzy neural network