if you keep it off ,the koordinates are still accurate of the last target but for some reason its now at -100 hight

The problem with storing the target data is that if you want switch targrt it will always snap back to the originl

Can't atm, maybe in a few weeks if I get time

Will you make a version with new composite?

also iirc lasers only work up to 2km right? Or has that changed?

yeah, but like if you use lasers you still have to point them using radar. It also introduces challenges like how do you know if the laser is pointing at the target or the terrain (Like say you're a radar above another flying target)

Since (to my knowledge) lasers do not have random noise applied to the found distance, there are simpler ways to compute this without the need for the Kalman filter

Can you also make it position tracking with a Laser that also can provide GPS coordinates

Just add me on discord @smithy3141 you can find me on stormworks official server

gotta talk to you.

Iirc it does output acceleration

what would i need to do to also get the mass and the xyz acceleration of the target?

Just this alone, only a few hours. A lot of this was just repurposed from other projects I had tbh. Like I learnt how kalman filters worked while making the jpdaf tws. But yeah a pretty long time if you consider all the research I had to do for other projects.

Neat! How long did it take you to make?

Target size can only be calculated with the old radars, but I could add range at some point

It would be nice to add a target size estimator to try to make the system more dynamic and versatile

good work, functional and efficient

Btw just throwing this out here, if you want to track a very manoeuvrable target increase the process noise a bit

velocity is in m/s and acceleration is in m/s/s

slight question what units are the speeds in ?

There’s a table somewhere that gives the max sweep speed for a given x fov, but I don’t have it on hand

If the sweep speed is too high and the range is too far, it’s possible for SW radars to completely miss targets

Oh you asking why that happens on your own build?

nono im just adding on top of what was said earlier

just target tracking

are these search radars spinning, because I haven't designed this to work with spinning radars at all

Yeah i've noticed that with search radars on my 2s6, when rotating it only detects for 4 - 5 times, then doesn't for 4 - 5 times again, repeats that all the time

I've been looking for something like this for awhile, thanks!

oh my god this is exactly what i was looking for thanks

amazing

ah well it was just like a couple vids that showed how SW radars don't update with new data every tick. With more range the the delay gets bigger too

im banned on SW discord

1. currently probs not. Mainly because of the weird time since detection mechanics. Like two identical radars may update out of sync, and also if the ranges are different they'll update at different rates.

if you search through some of my recent messages on the SWO discord for "time since detection", you can see what I'm talking about.

To get it working with the Kalman filter would require the data is like fuzed together somehow, which isn't what this kalman was designed for specifically.

no, im saying can i use it on multiple radars at once?

I'm not really sure what you mean by that. Like can it be used on the TWS? It's possible, but I'm not exactly sure how I'd implement that yet. It'd also be pretty performance heavy.

can the Unscented kalman Filter macro used on multi radar at once?

yeah very possible, you'll likely have to swap y&z composite outputs though (and then swap composite 2 and 3 in the other microcontroller).

is it possible for this to feed target data to a missile?

very nice!!

thank you

tom if you're asking if the system can be used with other radars, then yes. If you're asking if I use this in any of my other systems, then no. I only use a normal linear kalman filter for my V2 tws.

Added everything you wanted Nick. Also changed it so that you don't have to manually input time since detection now. Reworked how track initiation works too, so it'll auto select the target closest to the centre of the screen (and then try to keep this same target locked)

is the Unscented Kalman Filter used on other radar?

Alright, hyped for that !

Alright, I'll change the way the switch works when I change the way target selection works (so that it picks the target at the centre not the target in the first channel). I originally had a different idea to how the target selection would work, and that internal switch is just left over from that.

Tick compensation will be an easy add too. Already have target velocity and acceleration so I can just move the target position however many ticks into the future.

I could add a button to save target coords and another to swap Y&Z.

This will all happen sometime tomorrow though.

Nice system,
tho what could be added to make it more versatile since it will need to be modified alot of times to fit systems:
tick compensation
Save XYZ Or Not When target is lost, Swap Y&Z
Switch the toggle to a push, and instead use a push to toggle in the controller itself, so it's easily modified.
Description of each parameters and their effects.

time since detection is 0 for any radar with a max range less than like 2999m I think. Also that tsd calculator thing in the MC only gives a value if the radar has a target in it's sights.

don't quite get it, time since detection stays 0 when putting my own radar, is it normal ?