KOS Cruise Missile

[waltms]

My most recent attempt at an automated kOS cruise missile that targets hard-coded Long-Lat coordinates.

https://www.youtube.com/watch?v=Y1TPhWpTUEM

 

[Dman979]

That's pretty awesome! Nice stuff! I particularly liked the terrain-following feature, that was a nice touch.

 

It looked like it was yawing around a lot. Any ideas how to fix that?

[waltms]

12 hours ago, Dman979 said:

That's pretty awesome! Nice stuff! I particularly liked the terrain-following feature, that was a nice touch.

 

It looked like it was yawing around a lot. Any ideas how to fix that?

Thanks! The yawing towards the end of the flight is due to my script over-correcting and not being able to get out of it, so more work needs to be done, but it seems that in most flights, despite the swinging back and forth, it gets on target. I might even preserve the bug as a "feature" and call it an evasive maneuver, lol. I may just in the end put in an inline stabilizer, since with one it flies significantly better, but I've been trying to control flight with just aerodynamic surfaces.

[Dman979]

6 hours ago, waltms said:

Thanks! The yawing towards the end of the flight is due to my script over-correcting and not being able to get out of it, so more work needs to be done, but it seems that in most flights, despite the swinging back and forth, it gets on target. I might even preserve the bug as a "feature" and call it an evasive maneuver, lol. I may just in the end put in an inline stabilizer, since with one it flies significantly better, but I've been trying to control flight with just aerodynamic surfaces.

So it's a script issue, not an aerodynamics issue? It seemed to me like playing with the control surfaces might fix it, or having it bank into turns instead of yawing.

[waltms]

40 minutes ago, Dman979 said:

So it's a script issue, not an aerodynamics issue? It seemed to me like playing with the control surfaces might fix it, or having it bank into turns instead of yawing.

It might be both. I was allowing banking into turns before, but then sometimes the missile would slowly, but inevitably, go off in a crazy direction. My script is using both the PID cooked controls in kOS and some small adjustments using raw controls, and I think there is some conflict between the two that results in some of the current weirdness. What I will next try to do is tune the cooked controls and remove the raw control adjustments to see if that improves things. I will also make some adjustments to the surfaces, as you suggest, although I would like to keep it looking sort of like a Tomahawk if possible

[RoadRunnerAerospace]

that is officially amazing

 

[waltms]

3 minutes ago, RoadRunnerAerospace said:

that is officially amazing

 

Thanks!

[RoadRunnerAerospace]

I tried kos, and gave up

[waltms]

2 minutes ago, RoadRunnerAerospace said:

I tried kos, and gave up

I've put in more hundreds of hours into using kOS than I'd like to admit! 

[waltms]

3 hours ago, Dman979 said:

So it's a script issue, not an aerodynamics issue? It seemed to me like playing with the control surfaces might fix it, or having it bank into turns instead of yawing.

Dman,

Your instincts were totally right in terms of banking during turns. In my latest tests where I added banking logic the missile performs much better. It seems almost obvious, but when you're stuck in the minutiae of kOS logic it can be hard to see the forest for the trees. There will be a new video forthcoming.

[Dman979]

Just now, waltms said:

Dman,

Your instincts were totally right in terms of banking during turns. In my latest tests where I added banking logic the missile performs much better. It seems almost obvious, but when you're stuck in the minutiae of kOS logic it can be hard to see the forest for the trees. There will be a new video forthcoming.

Glad I could help!

 

[waltms]

Improved Avionics! Thanks Dman!

 

[RoadRunnerAerospace]

sooo how about throwing on a BDA warhead on that sucker? 

[waltms]

6 hours ago, RoadRunnerAerospace said:

sooo how about throwing on a BDA warhead on that sucker? 

An earlier version did have one, but in this version--it's a little difficult to notice--when it reaches within 500m of the target the fairing pops and a 1000lb Aviator Arsenal mod bomb is ejected and takes out the target. That's why the engine half suddenly tumbles out of control since the center of gravity suddenly dramatically shifts. I will likely play around more with the warhead configuration now that the flight script seems to be working sufficiently well. There is something unsatisfying about the current system since I think I'd rather see the entire missile slam into the target. What I'd really like to do is somehow manage an air burst a few meters over the target!

[RoadRunnerAerospace]

that would be pretty cool

[waltms]

Tuning to elevator controls for more stable flight, which also gives it more nimbleness in avoiding sudden terrain changes. The logic is generalized and not specific for this missile as it adjusts and learns which default elevator angle results in stable flight for the given craft's center of gravity and speed. The missile itself is different from the previous video, with different warhead configuration, length and center of gravity, 

 

 

[RoadRunnerAerospace]

Have you thought about adapting that program so that it does "search patterns", and use it as a science probe?

[waltms]

1 hour ago, RoadRunnerAerospace said:

Have you thought about adapting that program so that it does "search patterns", and use it as a science probe?

No, I have not. Hmm, so you mean like set it to zigzag across an area and at specific points do some science?

[RoadRunnerAerospace]

ye, for planets like eve and stuff

[waltms]

56 minutes ago, RoadRunnerAerospace said:

ye, for planets like eve and stuff

I have not thought about that, but will give it some thought.