Fault-tolerant Attitude Control of Sattelite Using Neural Networks

Sepehr, Firoozeh | 2014

563 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 46860 (45)
  4. University: Sharif University of Technology
  5. Department: Aerospace Engineering
  6. Advisor(s): Asadian, Nima
  7. Abstract:
  8. This research aims to introduce a new integrated fault tolerant attitude control system of satellites based on the neural networks. First, a linear controller and an optimal controller are designed assuming that no fault occurs in the reaction wheels of the satellite. It is obvious that these controllers are not able to respond properly when a fault occurs. Therefore, artificial neural networks are employed in the fault tolerant controller design. Neural networks are well-known for their adaptiveness and the ability to learn the dynamics of a system. Therefore, they can be used to predict the faulty conditions. In this research two approaches are examined to deal with the faulty conditions. In the first approach, neural networks are used to approximate the fault parameters. The estimations are used to correct the linear control law to counteract the actuator fault effects. In the second approach, neural networks are used as a part of the control loop to compute the control signal for compensation of the effects of the actuator faults. In the latter case, the control signal is the direct output of the neural network. In this work, different structures of networks are implemented and tested to discover their potential for implementation of the fault tolerant satellite attitude controllers
  9. Keywords:
  10. Neural Network ; Operator ; Active Control ; Fault-tolerant Control ; Satellite Attitude Control

 Digital Object List


...see more