Diversified Basis Functions-Based Fault-Tolerant Control for Electromagnetic Levitation System

The electromagnetic levitation system of maglev trains may encounter actuator failures during prolonged operation. Concurrently, the model of the electromagnetic levitation can deviate to a certain extent over time, leading to modeling uncertainties. These problems present formidable challenges to the control of the electromagnetic levitation. In light of the neural network (NN)’s capacity to approximate arbitrary nonlinear functions and given the presence of diverse nonlinearities inherent in electromagnetic levitation control, this study introduces a neuroadaptive fault-tolerant control algorithm that leverages diversified basis functions. The proposed approach demonstrates robustness against modeling uncertainties and exhibits fault tolerance towards actuator failures. A simulation analysis conducted in MATLAB/Simulink corroborates the efficacy of the proposed method.