3 ms·
>> and then it uses IR and radar together to remove stealth. So what happens when the dual seekers give different results? If there were three seekers then yo
by sandworm101 5y ago
>> and then it uses IR and radar together to remove stealth.
So what happens when the dual seekers give different results? If there were three seekers then you could ignore one, but with only two you cannot check one against another. What if the radar says turn left but the IR says turn right? Having two seekers may make the target's life easier: defeat either seeker and you defeat both.
- nradov 5y agoIt's like any other sensor fusion problem. You take a weighted average of the detected target bearing based on whatever data you have. Even if the target manages to completely "defeat" one sensor by somehow becoming invisible it doesn't mean that sensor will report a totally spurious bearing; you just won't get any meaningful data at all from it and thus ignore it.
- sandworm101 5y agoLeft + Right /2 = Strait = Miss. Small errors count. A few degrees off and your missile doesn't get to the target. Any averaging will probably result in a miss. In such cases a single sensor would be better, or two different sensors on two different missiles. This comes up in all manner of flight systems. Take altitude. If one sensor says you are at 1000feet and another says 10,000, the one thing everyone should agree is that you aren't at 5500. One of the sensors is wrong and needs to be ignored, but how do you pick between the two?
- nradov 5y agoThat's not how modern sensor fusion algorithms work. It's never a simple average calculation. The inputs from various sensors are weighted based on confidence levels. In extreme cases sensors reporting bad or inconsistent data are ignored.
- travisjungroth 5y ago> One of the sensors is wrong and needs to be ignored, but how do you pick between the two? What did the sensors say one minute ago? There's no reason this has to be Markovian.
- sandworm101 5y agoBecause, in a jamming/deception scenario, the jammer acting on a sensor won't just introduce a false reading. It will gradually transition from a true to false state in a manner that looks natural. The jammer will attempt to capture the sensor's attention and then shift its attention away from its target ("seduction"). In such cases it is very difficult/impossible to tell which is true based on past readings from the sensor. Google jamming and "capturing the gate" or "doppler pull", techniques specifically to defeat countermeasures dependent on past readings. With two sensors against a jamming target, the sensor with the clearest signal/most confidence is probably the one being most jammed. (IR flares do similar things too.)
- robotresearcher 5y agoThe parent said weighted average. The weight can be picked cleverly: it does not have to be the 0.5 of your example, and it doesn’t have to be constant. For example, the Kalman Filter tells you the weight to apply to each new sample and the last estimate to make a new estimate. These weights are optimal if the measurement errors have a Gaussian distribution with zero mean. In many practical systems they are usably good even when that assumption does not perfectly hold. In your example the new measurement that was less consistent with the last fused estimate would be downweighted by a KF.
- sandworm101 5y ago>> they are usably good even when that assumption does not perfectly hold. And when deployed against a jamming target that it well aware of gaussian distributions and how to game them? When deployed against a jammer that is very good at pretending to not be a jammer? An aware jammer will feed you good results, just enough to convince your gaussian distribution (capture) and them move it off of onto a false path (seduction) resulting in you following a false signal (miss) or cause you to fall back into search mode (delay=miss). The game is different when the thing you are trying to measure is actively trying to deceive you.
- robotresearcher 5y agoYou are right. Countermeasures would deliberately violate the assumptions a KF or other estimator makes, to degrade its ability to track. Now you have an arms race between countermeasures and the ability to model and defeat them.
- krona 5y agoThe Kalman filter is but one example of estimating unknown variable (the target location, heading, speed) using a joint probability distribution of multiple sensor inputs with different characteristics.