Parameterized Nonlinear Model Predictive Control Applied to Satellite Attitude Control

RODRIGUES, REURISON SILVA; MURILO, ANDRE; LOPES, RENATO VILELA; SOUZA, LUIZ CARLOS GADELHA DE

Lisboa

Conferencia; International Conference on Structural Engineering Dynamics; 2019

Instituto Superior Técnico Universidad de Lisboa e Instituto de Engenharia Mecânica (IDMEC)

The satellite attitude control and determination system (ADCS) is responsible for keeping it in position, speed and correct trajectories, stabilizing the spacecraft and guiding in the desired directions during the mission, regardless of external disturbances. Because satellites are devices that have physical limits, whether of actuators, sensors or structural, it is important to take these aspects into consideration in the design of controllers. A control strategy capable of dealing with constraints inherent in a physical system in a systematic way is the predictive control based on nonlinear models (NMPC). This strategy consists of solving a sliding horizon optimization problem, based on feedback information from the sensors. A major disadvantage of using NMPC is the high computational cost associated with its online optimization step. One of the techniques used to deal with this problem is the use of command parameterization. With this technique, it is possible to decrease the dimensionality of the original optimization problem. This ends up reducing the time spent in the calculation of the optimal control solution, enabling the implementation of NMPC in real-time environment. Once the control strategy has been developed, it should be tested and validated experimentally embedded in hardware. However, given the impossibility of usingthe real satellite for testing, it is necessary to use some methodology that allows the  validation of the controllers in a reliable environment. For this, two methodologies will be used. The first one is the Model-in-the-loop (MIL), aiming to test the controller only in virtual simulation environment like Matlab and the second one is Processor-in-the-loop  (PIL), to test the controller embedded in hardware. Thus, this work aims to develop a  parameterized embedded non-linear model predictive control (NMPC) for an air bearing satellite platform and, therefore, to validate the control strategy using MIL and PIL techniques.