A novel nonlinear model predictive control strategy and its application to maglev ball system
Tianbo Peng,Hongliang Li,Hui Peng,Xiaoying Tian,Yemei Qin,Xiaoyan Peng
DOI: https://doi.org/10.1080/00207179.2024.2366435
IF: 2.102
2024-06-20
International Journal of Control
Abstract:For processes with strong nonlinearity and fast response, a Nonlinear Model Predictive Control (NMPC) algorithm is proposed based on Laguerre functions and Radial Basis Function-based AutoRegressive model with eXogenous variable (RBF-ARX-LMPC), which is built to capture the nonlinear dynamics of such process. Firstly, the RBF-ARX model is transformed into an extended Non-Minimal State-Space (NMSS) model in which integral action and set-point information are naturally contained to decrease steady-state error. Then, to reduce computational burden, the control variables of the NMPC is parameterised by Laguerre polynomials for dimensionality reduction. The proposed control strategy is applied to a maglev ball system and is compared with Proportional–Integral–Derivative (PID) controller, the RBF-ARX model-based Linear Quadratic Regulator (RBF-ARX-LQR), and the RBF-ARX model-based Model Predictive Control (RBF-ARX-MPC). The experimental results show that the proposed control strategy improves transient performance and computational efficiency.
automation & control systems