New Canadian Shipbuilding Strategy

[JMCanada]

6+6 or even 6+8 would be nice to replace the MCDVs.
6 light frigates capable not only for coastal defence but also to be deployed to NATO drills, combat piracy or sail along the RCDs. With some AAW defense (at least point defense), helo and hangar.
Plus 6-8 OPVs with low crew requirements, similar to the Vigilance, with MCM capabilities (using USV/UUV), and a working deck which might accommodate three to six 40 ton containers to even act like a small-crewed arsenal vessel.

 

[suffolkowner]

I think any larger surface ship the Navy gets should have a VLS for BMD missiles. A Layered system is the best defence against todays and future threats.
With the River Class getting the SPY 7 Radar and having the A-OTHR system in place on land plus the SPY on the east coast. . I think having ships with the capability to be part of the BMD defence layer makes sense. Should form a critical part of the system. The Corvettes don't need to have a fancy radar, just be linked to the other detection systems for launch.

There is a lot involved in the BMD mission. It is not simply a matter of slapping SM-3 cans into strike length VLS. There is a whole radar calibration exercise that will make your eyes water, plus the training a certification bill for the crew is lengthy and persihable.

whats the deal with BMD and ships? Is it about intercepting the missiles in a specific part of the flight path?

 

[SeaKingTacco]

whats the deal with BMD and ships? Is it about intercepting the missiles in a specific part of the flight path?

Obviously, I have no first hand experience, but apparently it is harder than it looks. That was relayed to me by people who have done it.

 

[GR66]

There is a lot involved in the BMD mission. It is not simply a matter of slapping SM-3 cans into strike length VLS. There is a whole radar calibration exercise that will make your eyes water, plus the training a certification bill for the crew is lengthy and persihable.

My understanding is that the "successful" SM-3 BMD intercepts have all been done with freshly calibrated radars operated primarily by contractor staff in ideal weather conditions against non-maneuvering target missiles and no decoy systems.

Intercepting real ICBMs in a realistic scenario with ships in standard readiness is likely to have a pretty low success rate. Frankly having a dozen CDC's with a handful of SM-3's is unlikely to have any significant impact in the face of a full-scale nuclear attack by Russia and/or China which could see thousands of maneuvering re-entry vehicles and accompanying decoys.

Perhaps they might be effective against a smaller scale attack by say North Korea or Iran, but I have trouble envisioning a scenario where either of those countries would waste their limited nuclear arsenals against Canada.

All of that to say that I don't think that BMD should be focused on as a primary capability for the CDC. The time, effort and money would be better spent on other capabilities in my opinion.

 

[Oldgateboatdriver]

Everything about intercepting a long range missile in a ballistic trajectory is harder than it looks. It is the proverbial "trying to shoot down a small caliber bullet with an even smaller caliber one".

Think about it this way: It takes months for NASA to calculate even fractionally small corrections of course for the ISS, any satellites or probes. Now you want to calculate the exact course of a ballistic missiles with only minutes of times and limited original info, then calculate the very precise and exact intercept course of your own ballistic smaller missile.

 

[Czech_pivo]

Everything about intercepting a long range missile in a ballistic trajectory is harder than it looks. It is the proverbial "trying to shoot down a small caliber bullet with an even smaller caliber one".

Think about it this way: It takes months for NASA to calculate even fractionally small corrections of course for the ISS, any satellites or probes. Now you want to calculate the exact course of a ballistic missiles with only minutes of times and limited original info, then calculate the very precise and exact intercept course of your own ballistic smaller missile.

Maybe this is an area that AI will enhance, improve or even take over in the future.

 

[Oldgateboatdriver]

I doubt it.

The math is known and unless someone comes up with something better than special relativity, the only thing that can enhance this area is better sensors to more quickly and more accurately detect, classify and track the target and more parallel computing power (not smarter computers) to calculate track and intercept faster and more accurately. All of these things are known already and thus, no improvement from AI would occur.

 

[Czech_pivo]

I doubt it.

The math is known and unless someone comes up with something better than special relativity, the only thing that can enhance this area is better sensors to more quickly and more accurately detect, classify and track the target and more parallel computing power (not smarter computers) to calculate track and intercept faster and more accurately. All of these things are known already and thus, no improvement from AI would occur.

AI would be used to expand the current realm of computer power, chips, sensors - not necessarily the math.

 

[SeaKingTacco]

AI would be used to expand the current realm of computer power, chips, sensors - not necessarily the math.

That is a meaningless sentence.

How?

 

[Czech_pivo]

That is a meaningless sentence.

How?

AI used to expand/improve the ability to create faster sensors, chips, computer power.

AI Chips: What They Are and Why They Matter | Center for Security and Emerging Technology

The success of modern AI techniques relies on computation on a scale unimaginable even a few years ago. What exactly are the AI chips powering the development and deployment of AI at scale and why are they essential? Saif M. Khan and Alexander Mann explain how these chips work, why they have...

cset.georgetown.edu

 

[GR66]

I doubt it.

The math is known and unless someone comes up with something better than special relativity, the only thing that can enhance this area is better sensors to more quickly and more accurately detect, classify and track the target and more parallel computing power (not smarter computers) to calculate track and intercept faster and more accurately. All of these things are known already and thus, no improvement from AI would occur.

And this is to track and intercept a single incoming missile. Now figure in thousands of incoming missiles along with decoys and multiple AD launchers of various BMD capabilities spread across the continent and the surrounding maritime domain (and Space if Golden Dome goes forward as planned) all needing to be coordinated in the same few minutes.

As Reagan found out with Star Wars, a comprehensive BMD system is a quick way to bankrupt a country and is unlikely to be effective enough to reliably defend your cities in any case.

 

[Oldgateboatdriver]

Anti-Ballistic Missile defense systems have been around almost as long as ballistic missiles have.

So much so that the first "arms limitation" treaty between the US and the USSR was the ABM Treaty, which limited each side to protecting only one site in their country (the USSR chose Moscow, the US chose North Dakota, to protect the Air Force part of the "triad" at Grand Forks AFB). They did not reach that agreement from the kindness of their heart, but because both sides realized such system at the scale required to protect the whole country would quite simply bankrupt both of them.

Reagan's "Star War" project at least put a lot of money into fundamental physics and chemistry research for the most part and didn't go much further than that into practical application before being caned as the Cold War ended.

 

[Dana381]

Everything about intercepting a long range missile in a ballistic trajectory is harder than it looks. It is the proverbial "trying to shoot down a small caliber bullet with an even smaller caliber one".

Think about it this way: It takes months for NASA to calculate even fractionally small corrections of course for the ISS, any satellites or probes. Now you want to calculate the exact course of a ballistic missiles with only minutes of times and limited original info, then calculate the very precise and exact intercept course of your own ballistic smaller missile.

The Iron dome has been doing this for years, I have no idea how they do it but there system even has time to determine appx where the rocket will land and determine if sending an intercept is prudent. I imagine the iron dome has even less time to make its calculations as it intercepts mainly short range rockets.

 

[Oldgateboatdriver]

as it intercepts mainly short range rockets.

Exactly: Short range, thus not going into space, not traveling at the huge velocities of space using ICBM and detectable right from the moment of being fired. Also, going into an area about the size of P.E.I. In fact, those missiles don't even reach what is considered "hypersonic" speeds, whereof ICBM's are way past such speeds. The Iron Dome is just a local air defense system expanded to be made into an area one - not even a theatre one.

 

[Navy_Pete]

AI used to expand/improve the ability to create faster sensors, chips, computer power.

AI Chips: What They Are and Why They Matter | Center for Security and Emerging Technology

The success of modern AI techniques relies on computation on a scale unimaginable even a few years ago. What exactly are the AI chips powering the development and deployment of AI at scale and why are they essential? Saif M. Khan and Alexander Mann explain how these chips work, why they have...

cset.georgetown.edu

That article seemed to be about tailoring chips specifically to optimize them for use with AI, noting it has an ungodly computational demand and runs in at $100M a month.

That's doesn't mean they are necessarily better at general computing, just that they've been tailored to make them better at the kind of processing AI is using, because the scale of it makes sense to do that (same as people set up bitcoin mining machines, and there are different computer setups for office use, gaming, networking, etc).

Computer power isn't the limiting factor in targeting even 'basic' missiles with a straight arc from point to target, it's the accuracy of the sensors and the difference a 0.01 degree error on heading or a 1 m error on height/range will give on figuring out the target when you are talking about something running 1000s of kilometers. That's ignoring the fact that it's pretty straight forward to also alter the track and program routes or have flight paths, and you don't have to change it much to make a hit a big miss when you are talking about trying to anticipate where something is going to be that far ahead of time.

Even at relatively short range more advanced missiles will do all kinds of changes to the pattern to make it hard to get a hard kill, and it's not like you can do the same with a city. And even if you do get a hard kill you can still have tonnes of hyper or super sonic debris coming your way.

Now add in loitering munitions, guided artillery, drones etc it's getting easier to kill people faster than it's we're developing effective countermeasures, and usually much more expensive to defend as well.

 

[Czech_pivo]

That article seemed to be about tailoring chips specifically to optimize them for use with AI, noting it has an ungodly computational demand and runs in at $100M a month.

That's doesn't mean they are necessarily better at general computing, just that they've been tailored to make them better at the kind of processing AI is using, because the scale of it makes sense to do that (same as people set up bitcoin mining machines, and there are different computer setups for office use, gaming, networking, etc).

Computer power isn't the limiting factor in targeting even 'basic' missiles with a straight arc from point to target, it's the accuracy of the sensors and the difference a 0.01 degree error on heading or a 1 m error on height/range will give on figuring out the target when you are talking about something running 1000s of kilometers. That's ignoring the fact that it's pretty straight forward to also alter the track and program routes or have flight paths, and you don't have to change it much to make a hit a big miss when you are talking about trying to anticipate where something is going to be that far ahead of time.

Even at relatively short range more advanced missiles will do all kinds of changes to the pattern to make it hard to get a hard kill, and it's not like you can do the same with a city. And even if you do get a hard kill you can still have tonnes of hyper or super sonic debris coming your way.

Now add in loitering munitions, guided artillery, drones etc it's getting easier to kill people faster than it's we're developing effective countermeasures, and usually much more expensive to defend as well.

I found this section below (from the article), of interest and thought that this was where the use of AI could allow for more calculations to occur in a shorter timeframe using less transistors to do so.


AI chips also have other, AI-optimized design features. These features dramatically accelerate the identical, predictable, independent calculations required by AI algorithms. They include executing a large number of calculations in parallel rather than sequentially, as in CPUs; calculating numbers with low precision in a way that successfully implements AI algorithms but reduces the number of transistors needed for the same calculation; speeding up memory access by, for example, storing an entire AI algorithm in a single AI chip; and using programming languages built specifically to efficiently translate AI computer code for execution on an AI chip. (Section V and Appendix B.)

 

[Nvlgzr]

Exactly: Short range, thus not going into space, not traveling at the huge velocities of space using ICBM and detectable right from the moment of being fired. Also, going into an area about the size of P.E.I. In fact, those missiles don't even reach what is considered "hypersonic" speeds, whereof ICBM's are way past such speeds. The Iron Dome is just a local air defense system expanded to be made into an area one - not even a theatre one.

As you said, ICBM’s are hypersonic in the Mach 20 to 24 region plus the speed of an interceptor ; it’s difficult to comprehend that in the blink of eye the missile has travelled from one side of lovely Scarborough to the other side. Now Trump wants his “Golden” dome because gold is so much nicer than Iron. He better put a call in to Mike Myers AKA Gold Member.

 

[Navy_Pete]

Reading that they give up precision for volume, which makes sense for AI but less so for for something that specifically requires precision.

Accurately predicting trajectory is hampered by unknowns, errors and dynamic changes over massive distances when you talk about trying to actually hit a projectile with another projectile, except the projectile might also be actively manuevering and very difficult to detect to being with. The basic calcs are pretty easy on paper though, and why that old king kong throwing bananas at each other worked on old computers, it's just when you add that much distance, time and other variables over an ICBM range it all falls apart.

 

[GR66]

Maybe....Maybe with a dramatic breakthrough in quantum computing you might get greater probability of intercepting incoming ICBMs but you still have the issue of cost differential.

[Warning: info is from Chatgpt so take with a grain of salt and details on highly classified systems are hard to get regardless]

The most reliable US system for attempting to intercept ICBMs is the Ground-Based Interceptor. The estimated cost of each of these missiles is $70-100 million dollars each. By comparison, a Russian RS-24 Yars ICBM has 4 x nuclear warheads as well as decoys and penetration aids and costs an estimated $15-20 million dollars each.

So let's be fairly conservative and assume Russia can deploy 4 x ICBMs with 16 x warheads (plus decoys) for the same cost as a single GBI. Let's be extremely generous and say that a breakthrough in quantum computing gives the GBIs a 100% kill rate. You'd still need to have 16 times more GBI's than the Russians have warheads to ensure that nothing gets through.

Due to the difference in complexity and cost it will always be cheaper to build more ICBMs than the defender can afford to intercept.

 

[Underway]

Yay missile defence conversation. In a ship contect. First thing lets talk about the fire control problem.

You have to hit a moving object, with a moving projectile going through a moving medium, shot from a moving platform. That's multiple third order equations stacked on top of each other (angular acceleration, rotational acceleration, exact position calculations, predictive position calculations and so on).

To hit that moving object you have to detect, identify, track, localize and then target it before you can shoot it.

With a Ballistic missile a lot of these things are done more easily at different phases of its flight. Launch is pretty easy, it climbs and the math is straight forward. So this is why long range missiles are dangerous as you can protect them during the launch phase using distance.

In its ballistic phase its hard to track and localize as its moving very fast and the course keeps changing... you have to also try a predictive shot on where it might be, such as a crossing target (very hard to lead these sorts of missiles).

In its terminal phase IF you are the target AND you have high angle radar you can probably shoot it down as its flying directly at you. Lower altitudes restrict the speed of ballistic missiles due to how thick the air is, so Mach 5 is probably their max speed IF they are designed for it. But high angle radars are not normally a thing for ships, which is why missiles often have pop up attacks. The speed however is a real problem for a lot of radars as they cannot compute fast enough to get a targeting solution before its to late. A lot of anti ship ballistic missiles have terminal phases where they act more like normal supersonic cruise missiles, because hitting the moving ship is not easy either.

AI can help you learn to fight these (predictive modelling in wargames for example). But new AI isn't going give you a quantum leap in helping you fight the ship, there already are libraries of responses, with most combat management systems will automatically recommend when being attacked by something. Most good CMS can just fight the ship for you if time is an issue. AI on a ship may be able to learn the particulars of that ship over time which is interesting, so it makes better recommendations or optimizes sensors for the crew based on conditions or context.

But its not a magic bullet.