Generalized Predictive Control Based on Neurofuzzy Model for Electric Multiple Unit

In view of the complex, uncertain and nonlinear characteristics of the Electric Multiple Unit (EMU) operation process, the neurofuzzy model based on T-S fuzzy model is presented by data-driven modeling method. On the basis of the train traction characteristic curve and operation data, sub-tractive clustering is employed to ascertain the number of fuzzy rules, and the adaptive neurofuzzy inference system (ANFIS) is used to optimize the T-S fuzzy model parameters. The accuracy of the model is verified with China train control system level 3 (CTCS-3). Together with the neurofuzzy modeling, generalized predictive control (GPC) algorithm is designed to ensure high precision tracking control of train in both position and velocity. Simulation results show the effectiveness and validity of the method.