Satellites mostly rely on star trackers and gyros for precision attitude knowledge. Under impulsive thrust manoeuvres, I expect the star trackers to blur and provide no attitude knowledge. What is the strategy for precision attitude estimation in such a case? Open Loop gyro propagation? If you know the attitude during the manoeuvre, then you can use it as a fail-safe mechanism next to the accelerometer, for aborting thrust manoeuvre in the event of non-nominal rates, e.g. in case the thruster gets stuck.

  • $\begingroup$ What do you mean, specifically, by "Open Loop gyro propagation"? $\endgroup$ – Organic Marble Feb 2 '20 at 12:17
  • $\begingroup$ What amount of blur are you expecting, and what do you think would cause it? $\endgroup$ – user20636 Feb 2 '20 at 12:41
  • $\begingroup$ @OrganicMarble Attitude knowledge propagation based on gyro sensing only. So no sensor fusion between star tracker and gyro for example $\endgroup$ – Watch This Feb 2 '20 at 13:43
  • $\begingroup$ @JCRM, let's say that you have misalignment between the CoM and the thrust vector. It could be for a high slew manoeuvre or for thrust vector control of a satellite. If you are using small star trackers, exposure times are kinda important, which in turn is susceptible to blurring. Imagine night photography of a distant light source with a camera. $\endgroup$ – Watch This Feb 2 '20 at 13:48

This needs to be assessed on a case-by-case basis, with simulations, and with all hardware and algorithms properly modeled.

But to give an ideia of what is concerned and what is usual:

  1. Image blur on star trackers is a thing indeed, you may expect a star tracker to stop providing measurements while thrusting occurs.
  2. Some satellites such as geostationary ones in transfer orbit, which require long thrusting periods simply don't rely on star trackers, but rather on a coupling of gyros, sun-sensor and a spin stabilization strategy.
  3. For orbit maintenance, either in LEO or GEO, the duration of maneuvers can be very short, in the order of tens of milliseconds, such that one can afford to propagate attitude in open loop while maneuvering.
  4. Some activities might need to be interrupted during maneuvering, optical imaging for instance, telecom is not since it has less stringent requirements on pointing accuracy.
  5. Some agile satellites, specially those with control-moment gyros rely on gyro-based open loop attitude propagation while making fast maneuvers, as they expect star trackers to lose track due to angular rate alone.
  • $\begingroup$ Thanks! Had the same train of thought, a pre-calibrated gyro, under blinding conditions seems viable. But ofc, momentary blinding. Since eventually gyro-only performance won't match the gyro+star-tracker performance Coupling a precalibrated gyro with other sensors, apart from star trackers, might lead to increased noise and decreased performance. Unless it is being calibrated realtime with three-axis-magnetometer e.g. $\endgroup$ – Watch This Feb 3 '20 at 14:06
  • $\begingroup$ Magnetometers are not accurate enough to calibrate reasonably good gyros, AFAIK. To be clear, the measurement itself might be accurate, but actual field being measured (Earth's magnetic field) is polluted by several effects in the spacecraft, such as battery dipoles and electrical currents. $\endgroup$ – Mefitico Feb 3 '20 at 15:50
  • $\begingroup$ I agree, especially in smaller class of satellites, magnetometers pick up everything, even the spinning motion of reaction wheels. But if you stick them out with a boom, they kinda perform okayish. Since they are there in every satellite at least for the detumbling phase, you might as well consider using them to enhance the observability and nullspace measurement with regards to redundant IMUs on board. I am yet to see analytical work which proves this train of thought. But my intuition says that if you using a redundant set of gyros, then you calibrate them when the ST is available ... $\endgroup$ – Watch This Feb 4 '20 at 12:37
  • $\begingroup$ .. But when the star tracker (ST) is gone, you can possibly resolve some of the problems associated with null space observability of redundant set of gyros, by virtue of having access to some sort of attitude information via magnetometers. Take this all with a grain of salt though $\endgroup$ – Watch This Feb 4 '20 at 12:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.