Loading...

Neural Network Control Allocation of Satellites with Redundant CMG

Hosseinzadeh Salaty, Mohsen | 2015

430 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 47773 (08)
  4. University: Sharif University of Technology
  5. Department: Aerospace Engineering
  6. Advisor(s): Asadian, Nima
  7. Abstract:
  8. In this study the control allocation with failure tolerance capability in satellite attitude control using Control Moment Gyro (CMG) actuators is investigated. The pyramid configuration is utilized for arrangement of the actuators. Moreover, considering the singularity of the CMGs, controller is designed based on the angular velocity of the gimbals as the controller output, which prevents the singularity of the actuators. The controller is designed based on the optimal control approach and its cost function is the minimum control effort. The response of the system to the optimal controller in two cases of with or without failure of one actuator is investigated and is compared with the results of the linear control. The results of the designed optimal controller reveal that it can point the satellite in the commanded direction with less energy than the linear control in addition to avoiding the actuators from singularity conditions. However, the optimal control cannot be employed online because it is very time-consuming. Therefore, the optimal control is used offline to train a neural network. The neural network controller is then used online to control the attitude of the satellite. The Monte-Carlo simulation results demonstrate that the designed neural network controller is robust against the uncertain moments of inertia, the disturbance moments acting on the satellite, and the time delay of CMG actuators
  9. Keywords:
  10. Optimal Control ; Neural Network ; Control Moment Gyro (CMG) ; Control Configuration ; Fault Tolerance

 Digital Object List

 Bookmark

No TOC