Robust adaptive control based on RBF neural network for stochastic electromagnetic suspension system

Electromagnetic suspension systems are inherently unstable due to their highly nonlinear dynamics and sensitivity to stochastic noise and external disturbances. Designing a controller to achieve robust stability and high performance for such systems is challenging. This paper addresses this challenge by proposing a robust adaptive neural network controller for electromagnetic suspension systems with model uncertainty and stochastic disturbances. The controller utilizes a radial basis function neural network to approximate unknown system parameters and functions. The stochastic bounded stability theorem is employed to handle stochastic noise. Additionally, the command filter technique is employed to simplify the controller design by eliminating the need for continuous differentiation of virtual control signals. Furthermore, by integrating the command filter with the minimal learning parameter method, the computational burden associated with derivative and adaptive terms is significantly reduced. Simulation results demonstrate the effectiveness of the proposed controller in mitigating the effects of stochastic disturbances and model uncertainties, achieving robust stability and high performance.