I think a reasonably good model would be simply assuming that motor
chemistry is well mixed, and thus the Isp is fixed at launch, so it takes a
gaussian value centered on the manufacturers typical value with their
standard deviation, and wouldn't change over the flight (with the exception
of nozzle backpressure).
On 9 April 2016 at 18:18, Aaron Baker <bitmapverse@xxxxxxxxx> wrote:
I'm interested in using state-space tracking techniques to try to improve
position tracking for a rocket, but I need a process model for the motor.
I have the manufacturer thrust curve and a measure of variability, but
modeling the variability as noise doesn't really make sense because thrust
from moment to moment shouldn't be statistically independent of the
previous moment, it should be roughly the same and change only slowly.
What kind of model might be appropriate?
Aaron
