Iterative Learning Control Based on Neural Network and Its Application to Ni-Mn-Ga Alloy Actuator With Local Lipschitz Nonlinearity

• Published in: IEEE transactions on industrial informatics Vol. 20; no. 6; pp. 8138 - 8148
• Main Authors: Yu, Yewei, Zhang, Chen, Zhang, Xiuyu, Su, Chun-Yi, Zhou, Miaolei
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.06.2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Actuators; Adaptation models; Artificial neural networks; Convergence; Error analysis; Feedforward control; Hysteresis; iterative learning control (ILC); Learning; Magnetic hysteresis; Manganese; Metals; neural network (NN); Ni-Mn-Ga alloy actuator; Nickel base alloys; Nonlinearity; Tracking errors; Uncertainty
• ISSN: 1551-3203; 1941-0050
• DOI: 10.1109/TII.2024.3369229

The inherent hysteresis property of Ni-Mn-Ga alloy material is the main reason that affects the positioning accuracy of Ni-Mn-Ga alloy-based actuator. This study proposes an iterative learning control based on feedforward neural network (ILCBFNN) to eliminate the effect of hysteresis on actuator positioning accuracy. In addition, the convergence analysis problem of the system that is subject to system irreversibility, local Lipschitz nonlinearity, and iteration-dependent uncertainty, is investigated. Specifically, ILC is combined with the FNN to improve the adaptability and performance of the ILC. The global Lipschitz-like condition is established using the principles of mathematical induction and contraction mapping. Then, the convergence of the ILC process is analyzed by studying the variation of tracking error along the iteration axis. The obtained convergence condition ensures that the tracking error converges to a small region proportional to the initial state error. Experimental results verify the feasibility of proposed ILCBFNN method.