SerialCaveman

Monopulse Principles and Techniques? It seems like I was wrong and you were right. But, resolving correct target is valid for low number of targets (two target case was presented, and seems rather easy in theory), with ideal radar and with reduced tracking accuracy. With increasing amount of targets it becomes impossible, with 10-20% chance of breaking a lock completely off whole cluster of targets. However, it seems relatively easy to implement functionality, that recognizes multiple targets in beam and invalidates monopulse solution. That should prevent seeker antenna (and also a missile) from going haywire. I would rather expect it to go (literally) ballistic in direction of previously predicted impact point, until valid target is reacquired. Outside of this book, i found also few much fresher articles, that I had no time to thoroughly read and understand - but it seems that whole matter is more complicated than we might think.

I don't think that any jerking movements of missile seeker is realistic. As far as I know (not a radar specialist!), when presented with two- target problem, monopulse solution will tend to point somewhere between, not directly at one of them. In addition to, because seeker antenna angular rate is (gross simplification incoming) feed to autopilot as acceleration command, some kind of smoothing filter is added, so that noisy and erratic antenna movement doesn't translate into noisy and erratic missile trajectory - or at least, not as noisy and erratic. I think that SAHR would be even worse. ARH can use a little bit more precise data, combining target position from aircraft (quite precise if radar is in STT), and own seeker data, creating better target state estimation (at least, in theory). Not implemented in DCS, as we know. Still, as far as i know, monopulse have no ability to distinguish between multiple targets in beam, if these targets somehow are not filtered out earlier based on doppler or range gate. As far as I know, only AESA radars have such theoretical ability, and I'm not even sure if it is implemented on military radars.

I was talking only on scenario when missile is notched and target is releasing chaff - so that chaff stays in doppler gates for a long time. Then, after it went to chaff, there's no turning back - after few seconds target would have flown outside of missile beam, besides - missile would have moved its doppler gete to match chaff. At least if you track chaff, be good at it and don't go for random airplane! I think that when missile is not notched and can reject chaff based on doppler - it will, tending towards maximum closure rate for head - on, and minimum when tail-on, thus rejecting chaff in nearly 100% of cases. I also think that - in presence of Kalman, and also accurate enough on-board INS - it is no-brainer to make filter that takes both carrier aircraft AND on-board radar data, smashes it together, and provides target trajectory estimation more accurate than otherwise possible. Also it further increases ECCM capibility and reduces chance of switching to wrong target.

And I would fully agree. However, it is not the behaviour that we observe in game. Switch is not gradual, but abrupt and generates very sharp maneuver of the missile. I would slightly modify: If target closure below some value and chaff is there AND chaff is moving no slower than - for example - 50 m/s than aircraft, roll die and then snap radar to fake target. for x amount of seconds? Then no repeat, I don't think that missile would require. If it was given information from launching aircraft radar, it wouldn't have gone for chaff in the first place.

Of course, I wouldn't search for implemented Kalman in DCS code (except most probably F-14). I rather wanted to make a point, that rockets jumping from 300m/s airplane to 0 m/s chaff cloud is - to say at least - not realistic. Not only because of range and Doppler gating, but also thanks to kinematic filters - even when missile is notched. Hell, if i recall correctly, even R-73 and maybe even R-60M have analogue kinematic filters, that would try to reject flares based on motion characteristics. I also know, that radar simulation for missiles looks like it's non - existent in DCS. Target tracking works most probably like you described. As for quick fix to problem from this topic, I would be satisfied with simple check after roll die result told missile to go for chaff. If speed difference between tracked airplane and chosen chaff cloud is greater than - say - 50m/s, tracking stays on target. If it's less - switch to chaff. That should make missiles behave in more believable manner. EDIT: I've checked and I was wrong about IR missiles - kinematic filters were implemented later (eg. on Mistral), source: http://www.mediafire.com/file/nq6i1ja0ds9gqi4/Histoy_of_the_Electro-Optical_Guided_Missiles.pdf/file

First of all, hello to everybody, it's my first post on forum Wouldn't it be a no-brainer for AiM-120, at least in C variant, to implement some kind of Kalman filter? It should be then able to estimate target acceleration. Then, when monopulse solution would like to start tracking chaff instead of airplane, estimated target acceleration should fly throug the roof, and probably some kind of check should disregard current radar data as invalid. Then, with no valid radar data, Kalman target trajectory estimation woud be used in further tracking loop iterations not only for antenna pointing, but also for guidance, with gradually increasing size of doppler and range gates. That is, until valid target (not that far of estimated position) is detected, or else - radar returns to search mode.