MIMO Neural Models for a Twin-Rotor Platform: comparison Between Mathematical Simulations and Real Experiments

• Kerman Viana ^[1] ; Mikel Larrea ^[1] ; Eloy Irigoyen ^[1] ; Mikel Diez ^[1] ; Asier Zubizarreta ^[1]
  1. [1] Universidad del País Vasco/Euskal Herriko Unibertsitatea

     Universidad del País Vasco/Euskal Herriko Unibertsitatea

     Leioa, España

     
• Localización: 15th International Conference on Soft Computing Models in Industrial and Environmental Applications (SOCO 2020): Burgos, Spain ; September 2020 / coord. por Álvaro Herrero Cosío, Carlos Cambra Baseca, Daniel Urda Muñoz, Javier Sedano Franco, Héctor Quintián Pardo, Emilio Santiago Corchado Rodríguez, 2021, ISBN 978-3-030-57802-2, págs. 407-417
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • This work presents a neural model developed for a multivariable system with complex nonlinear dynamics, obtained through a tight methodology used both in simulation and in the real platform. In addition, this neural model has been studied and designed to meet the requirements of a predictive control strategy. A Twin-Rotor platform is used as an example of a Multi-Input Multi-Output (MIMO) system to study and analyse how a neural network is able to reproduce its nonlinear coupled dynamics and accurately estimate future system outputs. An in-depth study of the neural structures and their performance in the prediction of future states has been developed. Results show with comparisons, the modelization inaccuracies that appears when a proposal made just on the basis of a mathematical simulation is used to conclude the good performance of these MIMO neural models.