I have a suggestion that could make blind sweep MUCH more effective. Couldn't you use several square-shaped platings arranged in a dome-ish shape in order to provide more consistent detection? Having multiple sensory planes (surfaces) would increase the overall saturation of your airspace by rays. This does some nifty things for your array: Firstly, because you'd have your cameras spaced out along a three dimensional, spheroid surface instead of a single two-dimensional plane, you don't have to rotate your scanner as fast to achieve 360 coverage. Secondly, for the same number of cameras you have on this planar array, spacing them out along a spheroid plane covers far more area at once at far greater efficiency and far greater consistency over time, for a comparatively marginal increase in the amount of space it takes up. Lastly, is the combined benefits of being able to rotate your array slower and more efficiently saturating the airspace with the cameras' rays. Rotating your array slower reduces the speed at which the distant endpoints of each ray are travel through the covered airspace, due to something called *angular velocity*, which is how fast something is rotating relative to another point, in this case the axis your array rotates on. The more you reduce this angular velocity, the greater you will increase the chance your raycast pulses will have to contact an object before the cameras are no longer facing the direction of that target, up until the rotational velocity of the array is equal to the linear velocity of the object. Mathematically, the contact (detection success) rate at a specified distance from each camera is a parabola where the resulting percentage is calculated based on the difference between rotational camera ray velocity, the linear velocity of the contact, and the total volume of space occupied by that object (it's size). The lower the difference, and the larger the object, the higher the success rate. So to put this more simply, the chance of detection for a given camera on a rotating array is greater at a specified distance the closer the linear speed of the object and the rotational speed of the camera's ray are to one another at that distance. If the object is either moving faster or slower than the ray beam would be rotating at that distance, you will be less likely to lock onto that target, and that detrimental effect only increases as the speed difference increases between target and raydar beam. By increasing the number of detection planes in a radius, you increase the number of opportunities a ray has to connect with your target (saturation) and by reducing rotational velocity, you decrease the gaps in your airspace between each pulse within the raydar's maximum range. TL;DR, arrange your cameras in a dome, reduce the rotation speed of your radar array, and add more cameras to cover more degrees of your airspace. You will be able to detect smaller and smaller objects at exponentially greater distances the more effectively you saturate your airspace with raydar beams.
The cameras are able to raycast in a 90 degree cone, so there isn't much need to for a physical dome structure. Also, the dish would not be able to use all the cameras for tracking a locked target if they were not all facing the same direction. The rotation speed doesn't really matter, since the dish is always going to face any given direction a fixed portion of the rotation cycle, ie. faster spin means less time on a given direction, but the dish will circle back more often. In fact, the scanning will be more consistent if the dish spins faster, since slower spin means the target has more time to get closer while the dish is facing the other direction. Also, there's no guarantee that the target will be traveling in the same direction that the dish is spinning. If you actually do the math, it turns out that blindly shooting lasers to the sky and hoping that they happen to hit a target before it hits you is a hilariously ineffective detection method, even with a large amount of cameras. Lets say that you want to reliably detect a target with 10 square meter cross section at 800 meters away. The total surface area you have to keep track of is a half-sphere with a radius of 800 meters, which is about 4 000 000 square meters, which means that there are around 400 000 possible, distinct positions where the target can be at any given moment. A camera can test an 800 meter ray every 0.4 seconds. For a single camera, it would take almost 2 whole DAYS to sweep the entire tracking space. In order to do a complete sweep ever 5 seconds, you would need a raydar dish with 32 000 cameras on it! Or, you you know, you could just build a single interior turret, lol.
@@laihela Couldn't you take advantage of overlap with a circular arrangement of cameras to minimize gaps in coverage? The idea for this style of passive radar isn't to cover every possible angle of a given airspace but rather, to cover the most likely ones in which you'll encounter a contact based on the terrain around you, and the more I think about it, the more I think a central dome should be supplemented with several planar dishes. Also, ten square meters of cross section would be a rather small aircraft wouldn't it, assuming small ship blocks are approximately one square meter? That'd be an extremely lightly armed and armored craft wouldn't it, assuming it was even built for combat? Active radar would be better for detecting smaller craft to begin with, anyway. Could you improve tracking reliability at longer ranges and variety of angles by having multiple perimeter dishes that will sweep blindly while there's no contact, but will specifically raycast in the direction of the contact if a contact is established? I get the sense that having redundancy and backup systems would reduce the camera count needed for reliable coverage. In order to prevent confusion in targeting in this case, would it be possible to have the script prioritize data chronologically? That is to say, the first radar to make contact is prioritized by all other systems until it loses its tracking, at which point the next radar down the list to have contact is used, so on and so forth, until all systems have lost target lock? Thanks for your reply by the way, my post was largely a hypothesis and I enjoy the opportunity to learn from this discussion.
@@carlodagunz I'm not sure what you mean by passive radar in this context, since raydars are an active detection method by definition. Also, I don't understand what do you mean by "advantage of overlap with circular arrangement"? The raydar already does cast directly at the target (and remotely shares/prioritises targets between multiple dishes) once it has been detected. Tracking a know target is the easy part, the issue is detecting it in the first place. Each camera sweeps a certain 'volume' of air per unit of time, there's no way around this limitation. In order to track an X amout of airspace, with X resolution and X frequency, you need X amount of cameras. Look, even if we are generous and decrease the resolution area to 100 square meters (you can already fit 100 small rocket launchers inside a 10 square meter area, which I think is plenty), you would still need 3600 cameras to sweep a dome every 5 seconds at _800 meters_ distance. You might as well just use a single large grid turret, since they can detect targets of _any size_ almost _instantaneously_ at the _same range._ Your idea of only casting in directions with high probability of incoming targets is valid though, and would reduce the amount of tracking space required. However, this approach would introduce blind spots into the airspace, which the enemy might exploit if they figure out that they exist. I also enjoy these kind of discussions, it is the way how we learn new things. :)
@@laihela Truth be told, I'm not sure what I meant by passive radar either. But what I mean by a circular arrangement is that because cameras have a 90 degree cone, you can space them out on a dome in such a way that you have no blindspot on that structure. Let's assume that Raydar detection within 1km is irrelevant, since large grid turrets fulfill that purpose, and instead focus on the problem of detecting contacts at a distance of 2km or greater, where a number of engineered weapons (such as custom-made cruise missiles) are more of a threat than vanilla weaponry, especially turrets. I don't think the goal should be to achieve 100% coverage within 5 seconds because a ship is unlikely to ever be moving so fast that this is required. An aircraft flying directly through the center of your airspace, assuming a *minimum* diameter of 4km from any particular dish (since cameras can scan 2km in one second) and a velocity of 100 m/s (default speed limit), will take a minimum of 40 seconds to cross it, and 20 seconds to reach the center of that theatre, and that's just with a single dish.
You could combine this with a speedbreaking clang gun to strike from extreme range.
I have a suggestion that could make blind sweep MUCH more effective. Couldn't you use several square-shaped platings arranged in a dome-ish shape in order to provide more consistent detection? Having multiple sensory planes (surfaces) would increase the overall saturation of your airspace by rays. This does some nifty things for your array:
Firstly, because you'd have your cameras spaced out along a three dimensional, spheroid surface instead of a single two-dimensional plane, you don't have to rotate your scanner as fast to achieve 360 coverage.
Secondly, for the same number of cameras you have on this planar array, spacing them out along a spheroid plane covers far more area at once at far greater efficiency and far greater consistency over time, for a comparatively marginal increase in the amount of space it takes up.
Lastly, is the combined benefits of being able to rotate your array slower and more efficiently saturating the airspace with the cameras' rays. Rotating your array slower reduces the speed at which the distant endpoints of each ray are travel through the covered airspace, due to something called *angular velocity*, which is how fast something is rotating relative to another point, in this case the axis your array rotates on. The more you reduce this angular velocity, the greater you will increase the chance your raycast pulses will have to contact an object before the cameras are no longer facing the direction of that target, up until the rotational velocity of the array is equal to the linear velocity of the object. Mathematically, the contact (detection success) rate at a specified distance from each camera is a parabola where the resulting percentage is calculated based on the difference between rotational camera ray velocity, the linear velocity of the contact, and the total volume of space occupied by that object (it's size). The lower the difference, and the larger the object, the higher the success rate.
So to put this more simply, the chance of detection for a given camera on a rotating array is greater at a specified distance the closer the linear speed of the object and the rotational speed of the camera's ray are to one another at that distance. If the object is either moving faster or slower than the ray beam would be rotating at that distance, you will be less likely to lock onto that target, and that detrimental effect only increases as the speed difference increases between target and raydar beam. By increasing the number of detection planes in a radius, you increase the number of opportunities a ray has to connect with your target (saturation) and by reducing rotational velocity, you decrease the gaps in your airspace between each pulse within the raydar's maximum range.
TL;DR, arrange your cameras in a dome, reduce the rotation speed of your radar array, and add more cameras to cover more degrees of your airspace. You will be able to detect smaller and smaller objects at exponentially greater distances the more effectively you saturate your airspace with raydar beams.
The cameras are able to raycast in a 90 degree cone, so there isn't much need to for a physical dome structure. Also, the dish would not be able to use all the cameras for tracking a locked target if they were not all facing the same direction.
The rotation speed doesn't really matter, since the dish is always going to face any given direction a fixed portion of the rotation cycle, ie. faster spin means less time on a given direction, but the dish will circle back more often. In fact, the scanning will be more consistent if the dish spins faster, since slower spin means the target has more time to get closer while the dish is facing the other direction. Also, there's no guarantee that the target will be traveling in the same direction that the dish is spinning.
If you actually do the math, it turns out that blindly shooting lasers to the sky and hoping that they happen to hit a target before it hits you is a hilariously ineffective detection method, even with a large amount of cameras.
Lets say that you want to reliably detect a target with 10 square meter cross section at 800 meters away. The total surface area you have to keep track of is a half-sphere with a radius of 800 meters, which is about 4 000 000 square meters, which means that there are around 400 000 possible, distinct positions where the target can be at any given moment.
A camera can test an 800 meter ray every 0.4 seconds. For a single camera, it would take almost 2 whole DAYS to sweep the entire tracking space. In order to do a complete sweep ever 5 seconds, you would need a raydar dish with 32 000 cameras on it!
Or, you you know, you could just build a single interior turret, lol.
@@laihela Couldn't you take advantage of overlap with a circular arrangement of cameras to minimize gaps in coverage? The idea for this style of passive radar isn't to cover every possible angle of a given airspace but rather, to cover the most likely ones in which you'll encounter a contact based on the terrain around you, and the more I think about it, the more I think a central dome should be supplemented with several planar dishes. Also, ten square meters of cross section would be a rather small aircraft wouldn't it, assuming small ship blocks are approximately one square meter? That'd be an extremely lightly armed and armored craft wouldn't it, assuming it was even built for combat? Active radar would be better for detecting smaller craft to begin with, anyway.
Could you improve tracking reliability at longer ranges and variety of angles by having multiple perimeter dishes that will sweep blindly while there's no contact, but will specifically raycast in the direction of the contact if a contact is established? I get the sense that having redundancy and backup systems would reduce the camera count needed for reliable coverage.
In order to prevent confusion in targeting in this case, would it be possible to have the script prioritize data chronologically? That is to say, the first radar to make contact is prioritized by all other systems until it loses its tracking, at which point the next radar down the list to have contact is used, so on and so forth, until all systems have lost target lock?
Thanks for your reply by the way, my post was largely a hypothesis and I enjoy the opportunity to learn from this discussion.
@@carlodagunz
I'm not sure what you mean by passive radar in this context, since raydars are an active detection method by definition. Also, I don't understand what do you mean by "advantage of overlap with circular arrangement"? The raydar already does cast directly at the target (and remotely shares/prioritises targets between multiple dishes) once it has been detected. Tracking a know target is the easy part, the issue is detecting it in the first place.
Each camera sweeps a certain 'volume' of air per unit of time, there's no way around this limitation. In order to track an X amout of airspace, with X resolution and X frequency, you need X amount of cameras.
Look, even if we are generous and decrease the resolution area to 100 square meters (you can already fit 100 small rocket launchers inside a 10 square meter area, which I think is plenty), you would still need 3600 cameras to sweep a dome every 5 seconds at _800 meters_ distance. You might as well just use a single large grid turret, since they can detect targets of _any size_ almost _instantaneously_ at the _same range._
Your idea of only casting in directions with high probability of incoming targets is valid though, and would reduce the amount of tracking space required. However, this approach would introduce blind spots into the airspace, which the enemy might exploit if they figure out that they exist.
I also enjoy these kind of discussions, it is the way how we learn new things. :)
@@laihela Truth be told, I'm not sure what I meant by passive radar either.
But what I mean by a circular arrangement is that because cameras have a 90 degree cone, you can space them out on a dome in such a way that you have no blindspot on that structure.
Let's assume that Raydar detection within 1km is irrelevant, since large grid turrets fulfill that purpose, and instead focus on the problem of detecting contacts at a distance of 2km or greater, where a number of engineered weapons (such as custom-made cruise missiles) are more of a threat than vanilla weaponry, especially turrets.
I don't think the goal should be to achieve 100% coverage within 5 seconds because a ship is unlikely to ever be moving so fast that this is required. An aircraft flying directly through the center of your airspace, assuming a *minimum* diameter of 4km from any particular dish (since cameras can scan 2km in one second) and a velocity of 100 m/s (default speed limit), will take a minimum of 40 seconds to cross it, and 20 seconds to reach the center of that theatre, and that's just with a single dish.