Approximation capabilities of multilayer feedforward networks
Neural Networks
Modern control engineering (3rd ed.)
Modern control engineering (3rd ed.)
Fundamentals of Artificial Neural Networks
Fundamentals of Artificial Neural Networks
| Hi-index | 0.00 |
This paper is intended to present a neuro-PD (Proportional-Derivative) controller for tuning the dynamic response of a maglev vehicle running at high speeds around specific accelration amplitude. The maglev vehicle is simulated as a rigid car body supported by a rigid magnetic bogie-set with a uniformly distributed spring-dashpot system, in which the electromagnetic force is controlled by an on-board PD controller. Considering the motion-dependent nature of electromagnetic force working in a maglev system, this study presents an iterative approach to compute the dynamic response of the running maglev vehicle system based on the Newmark method. To determine the PD gains for a maglev vehicle traveling at various speeds, a proposed neuro-PD controller is trained using back propagation neural network (BPN) in such a way that its PD gains are correlated to the generated dataset of moving speeds and the maximum vertical accelerations of the maglev vehicle. Numerical simulations demonstrate that a trained neuro-PD controller has the ability to tune the acceleration amplitude of a running maglev vehicle within an allowable region of restricted acceleration.