I have some questions on how the uncertainty is computed for radar sensors. Since the twitch stream on satellite operations mentioned Kalman filters as a possible future addition, I played around with a singe radar to see what the uncertainty looks like, and here is the result:
The radar gets a first rough fix, then it gets refined on the 2nd and 3rd pings, then at the fourth ping it gets rough again.
What seems strange to me is that the resolution is not much better in the range direction than in the cross-range direction: I used a Nebo Uye radar which according to the db has a range resolution of 0.03 (nm I assume?) and a angular resolution of 1.2 (° probably), but that does correspond to what I see when running the scenario. Is that expected?
Note that if a very accurate range resolution were modelled, that would mean that with two radars you would get a very accurate fix on the position of a target by fusing the estimates, even if they have very poor angular resolution, and thus would be able to engage with SAMs since there would be no position ambiguity. That seems a bit too powerful for an old radar network, which would not be able to fuse the detections.