Why can't satellites be equipped with radar systems so that they can detect (and then avoid) impending collisions?

[dansmith170]

Why can't satellites be equipped with radar systems so that they can detect (and then avoid) impending collisions?

 

[DaveC426913]

Mass: satellites cram everyting they can into a tiny mass. Any extra is a waste of launch fuel.
Complexity: the satellite now needs a functional radar system as well as a fuel supply and propulsion unit. See above.

And the big one: Why? How often does this scenario actually come up?

Isn't it kind of like wearing a grizzly-proof suit when you visit the local park? I mean, the chances of getting attacked by a bear aren't zero, but surely it would be a better use of money to just make sure we keep bears out of our local parks?

 

[Baluncore]

Why can't satellites be equipped with radar systems so that they can detect (and then avoid) impending collisions?

The automatic collision avoidance radar on most satellites would not be needed before the satellite was out of date and out of fuel. Then, when a collision occurs, the collision avoidance radar becomes part of the fragments that must be avoided by other satellites.

If the weight of the collision avoidance radar, was made up instead by extra fuel, then the satellite could be controlled from the ground, to avoid the collision.

 

[mfb]

In low Earth orbit, most objects are closer to the ground than they are to other satellites. To catalog objects a ground-based radar system is much cheaper and much more powerful than a system you could put on a satellite.

A radar system looking in all directions for a possible imminent collisions would be far too heavy and power-consuming. It's too late as well. If you detect an object 10 km away that will come within 10 meters of your spacecraft two seconds later, what are you going to do? Spacecraft thrusters can't move out of the hazardous area that fast. You want to start avoidance maneuvers hours in advance, especially for satellites relying on ion thrusters (the majority of satellites now).

 

[Rive]

You can check the specs of radars used in fighter aircrafts for hints.
They require power in the range of tens of kilowatts (far more than available on a common satellite), and their weight is in the range of the smaller satellites.
Their price is also in the range of smaller satellites.
And for all that, their capabilities still fall short for the distances and speeds expected in space, even at LEO.

So - just does not worth it.

 

[dansmith170]

In low Earth orbit, most objects are closer to the ground than they are to other satellites. To catalog objects a ground-based radar system is much cheaper and much more powerful than a system you could put on a satellite.

Counterpoint: "The growth in the number of objects in space has created challenges for the Space Force. These include gaps in the geographical distribution of global sensors that collect data and limited sensor capability for objects in deep space." GAO-23-105565

Perhaps an onboard radar system may be appropriate for a larger satellite with greater power consumption that is beyond geosynchronous orbit (assuming the radar can detect a 10km/s collision in sufficient time).

 

[dansmith170]

And the big one: Why? How often does this scenario actually come up?

According to Space.com "Since the launch of the first Starlink spacecraft in 2019, the SpaceX satellites have been forced to move over 50,000 times to prevent collisions." And it appears that half those maneuvers were in the past 6 months. Does that answer your "why" question DaveC426913?

 

[Baluncore]

Perhaps an onboard radar system may be appropriate for a larger satellite with greater power consumption that is beyond geosynchronous orbit (assuming the radar can detect a 10km/s collision in sufficient time).

Why should a big satellite move out of the way of a small satellite, or fragment, if it was easier to swat, or deflect the small one?

 

[DaveC426913]

According to Space.com "Since the launch of the first Starlink spacecraft in 2019, the SpaceX satellites have been forced to move over 50,000 times to prevent collisions." And it appears that half those maneuvers were in the past 6 months. Does that answer your "why" question DaveC426913?

Ok, so I see the issue.

I thought you were concerned about whether or not we had the ability to track and avoid collisions.
We do; your concern is why not upload the tracking heavy-work to the satellites themselves, as opposed to where we have it now, which is ground-based?

And the answer, as others have pointed out, is that tracking potential collisions in real-time is impractical to the point of impossible. By the time a satellite could detect a potential collison, it is too close and can't move out of the way fast enough. Orbits are very predictable and any potential collisions can be seen well in advance, which means the logistics of it can be handled easily from the ground.Imagine it as similar to air traffic control near an airport. If every plane departing or arriving were responsible for maintaining its own clear airspace (which it would have to do in real time by radar) then there would be a lot more near misses, and a lot more planes hitting the brakes hard or hitting the gas to wildly dodge a near-miss (and sending it off course). Better to have the whole thing orchestrated by the tower, where they have better radar and the ability to see planes coming for scores of miles (minutes) instead just a few miles (seconds).

(Admittedly, it's a bit of a contrived analogy. For example, you'd have to imagine the radar and fuel required for planes was a prohibitive fraction of their mass/volume/cost.)

 

[DaveC426913]

Perhaps an onboard radar system may be appropriate for a larger satellite with greater power consumption that is beyond geosynchronous orbit (assuming the radar can detect a 10km/s collision in sufficient time).

Keep in mind what we know from our killer asteroid dodging studies.
A small devation well in advance is way better than a huge devation at the last second.

A satellite having to wait until a collison is only seconds away has to leap out of the way. That's a huge outlay of fuel and a huge deviaton from its orbit, whch it now has to undo with a second burn of the same size.

The more advanced notice you have, the less acceleration you need to get out of the way, the less you need to get back in position - and a LOT less fuel you need for both.

Unforgiving orbital mechanics - that harshest of mistresses - may be the single biggest showstopper as to why we don't leave collision-avoidance to the individual.

 

[dansmith170]

Ok, so I see the issue.

I thought you were concerned about whether or not we had the ability to track and avoid collisions.
We do; your concern is why not upload the tracking heavy-work to the satellites themselves, as opposed to where we have it now, which is ground-based?

And the answer, as others have pointed out, is that tracking potential collisions in real-time is impractical to the point of impossible. By the time a satellite could detect a potential collison, it is too close and can't move out of the way fast enough. Orbits are very predictable and any potential collisions can be seen well in advance, which means the logistics of it can be handled easily from the ground.Imagine it as similar to air traffic control near an airport. If every plane departing or arriving were responsible for maintaining its own clear airspace (which it would have to do in real time by radar) then there would be a lot more near misses, and a lot more planes hitting the brakes hard or hitting the gas to wildly dodge a near-miss (and sending it off course). Better to have the whole thing orchestrated by the tower, where they have better radar and the ability to see planes coming for scores of miles (minutes) instead just a few miles (seconds).

(Admittedly, it's a bit of a contrived analogy. For example, you'd have to imagine the radar and fuel required for planes was a prohibitive fraction of their mass/volume/cost.)

Thanks for your replies Dave. Part of what is motivating my question about onboard radar systems for satellites is that ground based detection is not as good for xGEO orbits; thus the need for some kind of space situational awareness (SSA) mechanism for such orbits.

Most satellites orbit at 7 km/s. A collision of 2 satellites moving perpendicular to one another would be about a 10 km/s collision speed. That means a radar with a range of about 300-600 km would provide about a minute of reaction time for a satellite to maneuver so that it avoids collision. A much better alternative to mission failure.

In my opinion, the question then becomes: can we build a radar system that is less than a ton, that uses kilowatts or less of average power consumption, and that has an effective range of about 400 km?

 

[dansmith170]

Why should a big satellite move out of the way of a small satellite, or fragment, if it was easier to swat, or deflect the small one?

What mechanism would you suggest for deflecting a smaller satellite? And would that even be legal? Would it be safe to go around deflecting satellites off their original course?

 

[DaveC426913]

...ground based detection is not as good for xGEO orbits...

How often do impending collisions come as a surprise?

I think were pretty good at knowing orbital characteristics and predicting potential collisions, as witnessed by the relatively low number of collisions that have occurred.

Again, if we know an event is coming long in advance (because we can pro-actively predict it, then the whole issue of detecting it and having to wait until the last minute - at which point the satellite effectively has to panic.This sort of feels like a solution in search of a problem.

I think you would do well to step back from the solution and ask yourself this:

Are unrepredictable potential collisions of such a real concern that our current method of tracking and projecting paths is inadequate and that we need to move from this pro-active tracking to just-in-time detection?
Again, like with airplanes, why deprecate a system that can see potential collisions a long time ahead for a system that has to wait until we can see the whites of their eyes before taking action?

 

[russ_watters]

Perhaps an onboard radar system may be appropriate for a larger satellite with greater power consumption that is beyond geosynchronous orbit (assuming the radar can detect a 10km/s collision in sufficient time).

@mfb 's point still stands except now instead of being hundreds of miles away from earth/each other, now they are thousands or tens of thousands of miles apart. And the number is much, much smaller.

starlink

Orbits at 340 miles.

 

[Vanadium 50]

The AN/APG-79 radar weighs maybe 300 pounds and an average satellite weighs a ton. So it will fit, but at considerable cost in weight. However, it also takes 5 kW of power, and a typical satellite has only 500 W available. Where do you plug it in?

I think you need to look at economics. There are 5000 satellites in LEO and there have been about 10 collisions. So it's a 0.2% effect. If I add radar, it's a 15% or so loss of payload mass. 15 >> 0.2.

 

[mfb]

Counterpoint: "The growth in the number of objects in space has created challenges for the Space Force. These include gaps in the geographical distribution of global sensors that collect data and limited sensor capability for objects in deep space." GAO-23-105565

Perhaps an onboard radar system may be appropriate for a larger satellite with greater power consumption that is beyond geosynchronous orbit (assuming the radar can detect a 10km/s collision in sufficient time).

That's not a counterpoint. It's an argument for more ground-based radar stations.

The largest collision risks occur in low Earth orbit. At GEO and beyond we are looking at less than 0.1% risk over the lifetime of a satellite. You want to double its power consumption and significantly increase its mass just to work on that tiny risk?

According to Space.com "Since the launch of the first Starlink spacecraft in 2019, the SpaceX satellites have been forced to move over 50,000 times to prevent collisions." And it appears that half those maneuvers were in the past 6 months. Does that answer your "why" question DaveC426913?

... and no collision happened. That shows how good the existing ground-based detection is.

In my opinion, the question then becomes: can we build a radar system that is less than a ton, that uses kilowatts or less of average power consumption, and that has an effective range of about 400 km?

Probably not, and even if we could we wouldn't put it on a satellite. If you see a range quoted for a radar system then it's looking for aircraft and similar sized objects.

We know all objects larger than 10 cm in low Earth orbit. Satellites avoid these routinely already. Detecting them with an on-board radar is pointless. Finding more smaller objects would be useful - but satellite-mounted systems would have an atrocious range for them, which makes them pointless.

 

[Rive]

Perhaps an onboard radar system may be appropriate for a larger satellite with greater power consumption that is beyond geosynchronous orbit (assuming the radar can detect a 10km/s collision in sufficient time).

You should consider the distances properly. For a ground based radar system to reach the most 'dense' LEO orbits is at most a few thousand km range. To do it from geosynchronous orbit, it's around ten times the range required.

Detecting collisions - well, since orbits can be calculated, it's about mapping and calculations and not supposed to be about low reaction time.

I won't say that there won't be orbital radars ... devices with possible radar options, ever.
But likely not primarily for collisions, and not in the near future.

 

[Vanadium 50]

According to Space.com "Since the launch of the first Starlink spacecraft in 2019, the SpaceX satellites have been forced to move over 50,000 times to prevent collisions."

I don't understand why you didn't continue on. From the same article: ""SpaceX currently conducts an avoidance maneuver every time orbital models show a probability higher than 1 in 100,000 that one of the Starlink satellites will cross another object's path. That threshold is 10 times lower than the standard upheld by NASA and other international agencies."

So, first, SpaceX is moving its satellites around like crazy. Second, the expected number if nothing is done is 0.5 collisions. Third, we have had no collisions. It's hard to say how much the motion helps, since the number is already too low to measure accurately.

This is in the engineering section. Adding 15% of the cost (a reduction of 15% in the payload weight requires 15% more satellites to do the job) is Bad Engineering. If this 15% is not obviously any better than the existing solution, it is Very Bad Engineering. If the system in place seem to largely work, and the proposed system is questionable, it is Very Very Bad Engineering.

Finally, 15% more launches means 30% more opportunities for collisions. This makes the problem worse, not better! And that is Very Very Very Bad Engineering.

 

[Bandersnatch]

See, you think you're thinking like an engineer, but you're not. Putting a radar on a satellite brings it one step closer to a full-fledged space fighter, and that is just so very very very cool.

 

[Vanadium 50]

Putting a radar on a satellite brings it one step closer to a full-fledged space fighter,

Well, there is the question of which satellite moves to avoid the collision. The answer seems to be "the one with the smaller laser".

 

[harsh_raj]

Putting a radar on a satellite sounds like a good idea. Let's say we put one on Hubble. Since radar cannot work on its own, it needs to be powered by some source, so let's include batteries for that. Now, to avoid collision, we need some propulsion system (mini rockets), so let's add that as well.

After all this, Hubble becomes twice its mass, and we don't have a rocket to launch such a big satellite. So, there you have your answer. We can always make things overly complicated, but it always comes down to feasibility in terms of usage and economy.

 

[mfb]

Essentially all satellites have power, almost always via photovoltaics. Batteries are used to make it through the shadow of Earth. Almost all satellites have propulsion to maintain their orbit and attitude.

The two things you want to add are already essential components of almost all spacecraft.

 

[marcusl]

Essentially all satellites have power, almost always via photovoltaics. Batteries are used to make it through the shadow of Earth. Almost all satellites have propulsion to maintain their orbit and attitude.

The two things you want to add are already essential components of almost all spacecraft.

I'm guessing you've never designed a satellite or a radar, either. Rive has the correct answer in post #5 above.

 

[sophiecentaur]

Unforgiving orbital mechanics - that harshest of mistresses - may be the single biggest showstopper as to why we don't leave collision-avoidance to the individual.

I think orbital mechanics has been ignored in most of the comments in the thread. The problem is not as easy as for flying in busy air space. The requirement for a 'radar' system is very much three dimensional. Forward - looking is not enough. You need to look several orbits in the future to decide on the risk of collision and, except for space junk, you need to negociate between craft owners about who gives way. (Nothing like as simple as deciding who is the stand-on vessel in the Dover Strait - an even there you get collisions).
The solution has to be Earth-based , both for detection / avoidance and political negotiations and it would be like a worldwide Air Traffic Control.