An improved neural network tracking control strategy for linear motor-driven inverted pendulum on a cart and experimental study

• Published in: Neural computing & applications Vol. 34; no. 7; pp. 5161 - 5168
• Main Authors: Ping, Zhaowu, Zhou, Mengya, Liu, Chenxi, Huang, Yunzhi, Yu, Ming, Lu, Jun-Guo
• Format: Journal Article
• Language: English
• Published: London Springer London 01.04.2022; Springer Nature B.V
• Subjects: Adaptive control; Artificial Intelligence; Compensation; Compensators; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Data Mining and Knowledge Discovery; Electric motors; Friction; Image Processing and Computer Vision; Neural networks; Pendulums; Probability and Statistics in Computer Science; Special Issue on Computational Intelligence-based Control and Estimation in Mechatronic Systems; Tracking control; Tracking problem; Inverted pendulum on a cart; Adaptive friction compensation; Neural network; Tracking control; Discrete-time nonlinear output regulation
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-021-05986-9

Much recently, based on the discrete-time nonlinear output regulation (NOR) theory, a neural network (NN) method combined with feedforward friction compensation was proposed to tackle the tracking problem of the linear motor-driven inverted pendulum on a cart (IPC) system, where the NN was used to approximate the solution of the discrete regulator equations (DREs) for the IPC system whose dimension is equal to the sum of the dimensions of the system state and the control input. However, it is quite tedious to calculate the approximate solution and the feedforward friction compensation requires a complicated off-line friction identification procedure. This paper proposes an improved NN method combined with adaptive friction compensation for this tracking problem. In particular, the NN is used to approximate a feedforward function instead of the solution of the DREs. Since the dimension of the feedforward function is the same as that of the control input, the computational efficiency is improved. Moreover, the adaptive friction compensator is easy to implement since it does not need to establish detailed friction model. Experimental results demonstrate that our control strategy can lead to much more satisfactory tracking performance compared with the existing NN control strategy.