MIMO decoupling control based on support vector machines αth-order inversion

FuHua Song,Ping Li
DOI: https://doi.org/10.1109/WCICA.2006.1712495
2006-01-01
Abstract:Combining inverse system method and support vector machines (SVM), the paper presents an αth-order inverse system method based on SVM for general nonlinear MIMO discrete systems. The nonlinear offline inverse model of the controlled plant is built by SVM with the radial basis function (RBF) kernel. Cascading SVM ath-order inversion with the original system can form a pseudo-linear system. So the traditional linear system methods can be used in a nonlinear system. Simulations demonstrate that even without prior information about the plant, the method presented can accurately model the inverse dynamic system of the plant. This method can apply to general nonlinear MIMO discrete systems, and can be easily implemented for engineering. © 2006 IEEE.
What problem does this paper attempt to address?