Dynamic Positioning of a ROV Using a NMPC-Based Controller

• Jose Joaquin Sainz ^[1] ; Elias Revestido Herrero ^[1] ; Jose Ramon Llata ^[1] ; Francisco Jesus Velasco ^[1]
  1. [1] Universidad de Cantabria

     Universidad de Cantabria

     Santander, España

     
• Localización: WWME 2022 IV. Jardunaldia - Berrikuntza eta irakaskuntza energia berriztagarrien aurrerapenetan / coord. por Aitor J. Garrido Hernández, Matilde Santos Peñas, Fares Mzoughi, Irfan Ahmad, Izaskun Garrido Hernandez, 2023, ISBN 978-84-1319-526-1, págs. 3-6
• Idioma: inglés
• Enlaces
  • Texto Completo Libro
• Resumen

  • In this work a control with two control loops is presented. In the first loop an integral action is implemented and in the second loop a controller based on a nonlinear model (NMPC) together with an Unscented Kalman filter (UKF). These have been used to perform the dynamic control of a remotely operated vehicle (ROV). Due to the characteristics of the ROV dynamics, this type of NMPC control has been selected because it is especially suitable for nonlinear systems. Different external agents that may affect the system have been taken into account. Such as the noise of the sensors present in the system, modeled as white Gaussian noise. In addition, the drift effects or second order effects of the waves that could cause great damage to the dynamic control have been taken into account. To estimate and filter the sensor noise, a UKF filter has been implemented. This type of filtering is usually used in the filtering of disturbances for this type of nonlinear systems. It is necessary to incorporate a filtering system to the control system because if the sensor signals were not implemented, they could not be fed back to the controller because it would not correctly interpret the oscillations present in them and we would have a malfunction of the controller. Finally, in order to determine the validity of the implemented controller and filter, simulations are performed and their results are analysed.