MechJeb2

MechJeb2

4M Downloads

Ascent Guidance miscalculates rendezvous time

EoD opened this issue ยท 9 comments

commented

When I press "Launch to Rendezvous", the launch is too late and the orbiting partner is long gone when I reach it. It does improve, when I increase "Orbit altitude" to a higher value.
image
image

I am using Mechjeb 2.7.0.

As there seem to be a (missing?) correlation between the orbit altitude and the rendezvous timer, would it be possible to auto-set the orbit altitude to something similar to the "target altitude/apoapsis"? This would avoid unnecessary fuel consumption.

commented

@sarbian I am a bit confused. I did launch the vessel and quickloaded (as Minmus has no spaceports), but it didn't work. What exactly do you mean?
As you can see I did enable the "Skip Circularization". Did this break MJ's prediction?

commented

Hum, I am not sure it would work when launching from anywhere. The recoreder may not start...
I'll have to check

commented

Can you explain to me (I have no clue), why it is not possible to calculate an estimated rendezvous by using the targets orbit (median of periapsis/apoapsis), the current position (or just the angle?) and time it would take to reach the orbit's height without overshooting?
Of course assuming that there is no atmosphere.

commented

Where did I say it is not possible ? The code is not there. If someone want to write it they are welcome to do it.
Currently we have a solution that work well for most needs (launch from a space center) and the rendez-vous autopilot for other cases.

commented

This function needs the vessel to be launched once untill circularisation for MJ to know the time to reach orbit . Then you do do a revert to launchpad and use the launch to rdv.

commented

Does the launch and revert have to be done every launch, or can an old/original number be reused for a similar launch vehicle?

commented

It saves the value for a vessel name.

commented

It is entirely possible but the problem of doing it without a prior rocket launch is pretty much a post-grad level trajectory optimization problem -- particularly the "without overshooting" bit. I've spent the past 6-8 months off and on trying to teach myself how to do it, and I took graduate level quantum mechanics in college...

commented

#1061 is the rewrite that will eventually be able to solve this right, closing as a dup of that.