An alternative approach for the model based control of a nonlinear dynamic system

Alexander Verde, José Canelón, Néstor Queipo


This paper presents an alternative approach for the model-based control of the ";ball and beam";, a multivariable nonlinear dynamic system, which is a benchmark for testing new control algorithms. The proposed strategy uses two neural networks and a polynomial interpolating scheme to construct the desired value trajectory. The performance of this strategy significantly outperforms the corresponding to the classic linear quadratic regulator. Both strategies were implemented on computer simulations of the system, and their performance was evaluated using the following criteria: rise time, settling time, overshoot percentage, integral of the error's absolute value, robustness and design easiness. The control strategies were tested under step and sinusoidal changes of the reference value.

Palabras clave

neural networks; LQR; model based control; nonlinear dynamic systems

Texto completo:


Universidad del Zulia /Venezuela/ Revista Técnica de la Facultad de Ingeniería/ /

p-ISSN: 0254-0770 / e-ISSN: 2477-9377 


Licencia de Creative Commons
Este obra está bajo una licencia de Creative Commons Reconocimiento-NoComercial-CompartirIgual 3.0 Unported.