Abstract: This paper focuses on the modeling of magnetic levitation control system using RBF-ARX model,also discuses the principium of RBF-ARX model,model structure design,model parameters identification,parameter optimization and so on.Finally,the authors have used different series as state variable,established the RBF-ARX model separately and analyzed the reliability and performance of the model respectively.The results of simulation studies finally show the effectiveness of RBF-ARX model for the modeling and identification of nonlinear systems.In contrast to ARX model,RBF-ARX model is good at modeling of nonlinear systems.

Key words: magnetic levitation systems, model-based predictive control, RBF-ARX model, Structured Nonlinear Parameter Optimization Method（SNPOM）

摘要： 采用一种新的ARX模型（RBF-ARX模型）对磁悬浮系统进行离线建模,讨论了RBF-ARX模型的原理、结构的选取、模型参数辨识和RBF参数优化等问题。文章分别采用不同的序列作为状态变量,分别建立RBF-ARX模型,并分析了各模型的性能及可靠性。模型的预测输出和仿真结果,证实了RBF-ARX模型在非线性系统建模和辨识中的有效性。通过与ARX模型的比较,证明了RBF-ARX模型在非线性系统建模中效果更好。

关键词: 磁悬浮系统, 模型预测控制, RBF-ARX模型, 结构化非线性参数优化策略