M.Sc Thesis


M.Sc StudentDov Verbin
SubjectTime Efficient Angular Steering Laws for Rigid Satellite
DepartmentDepartment of Aerospace Engineering
Supervisor Full Professors Ben-Asher Yoseph
Full Thesis textFull thesis text - English Version


Abstract

The control of rigid satellite rotational maneuvering has been addressed by various researchers since the sixties of the 20th century. In this frame, this work addresses the design of an efficient rotational maneuvering control for a rigid satellite that is done by use of reaction wheels with a time invariant closed loop controller.

The efficiency requirement refers mostly to the reduction of the maneuver time, but also to algorithm simplicity, avoidance of control chatter and compatibility to variety of initial conditions and end conditions. This requirement  is typical for cases in which the payload is strapped to the satellite body, and its orientation has to be shifted from one target to another. Attention is paid for the smooth and transition from maneuvering to steady state tracking.

The assumption of rigid satellite may be taken when the decay time of the mode with the lowest frequency is small enough comparing to the time constant of the attitude control system.

The Reaction Wheels ( RW ) are the most common actuators for rotational control of satellites since they are basically simple instruments and they do not consume propellant. They are the only actuation method that is considered in this work. It is assumed that at least three of them are installed in the satellite at linearly independent directions. It is also assumed that each reaction wheel has a torque limit, and may also have an angular momentum limit.

The time invariant closed loop controller means that no reference trajectory is available for the mid-time until reaching the required final attitude, and the control at any moment is calculated as a function of the current measured attitude and angular rate without any use of time dependent gains or functions.

    This work proposes a new solution to the above defined problem which offer some advantages over existing solutions such as :
  • Efficiency for general three dimensional rotation.
  • Simplicity of algorithms.
  • Covering all cases of four or three reaction wheels at general alignment
  • Smooth control without chattering.
  • Tuning capability to compensate for time delays and parasitic dynamics.
  • Tuning capability to trade between performance and robustness.
  • Address the major factors that limit the performance which are the controller bandwidth, angular acceleration and angular rate capability in 3 dimensions.

Simulation analysis results are presented in this work to demonstrate the validity and potential value of the method.