Wednesday, January 9, 2019

Outranging The Sensor - Long Lance Torpedo

We’ve repeatedly noted that a weapon that outranges its targeting sensors is useless.  The current classic example is the much-hyped Chinese DF-21D ‘carrier killer’ anti-ship ballistic missile.  The missile has a range of around a thousand miles but there are no Chinese sensor systems capable of providing targeting data at that distance.  Thus, the missile is useless as a carrier killer.  The weapon outranges the sensor.

Let’s take a look at a historical example of this phenomenon, the Japanese WWII Type 93 Long Lance torpedo.  The torpedo had a range of up to 25 miles at slower speeds of around 35 kts and was capable of 14 miles at around 50 kts!  On paper, this weapon should have allowed the Japanese to stand well off from the US Navy ships and utterly wipe them out but it failed to do so.  Certainly, the torpedo did do serious damage but fell well short of being the weapon that it could have been because the Japanese had no sensors capable of detecting targets at ranges equal to the range of the weapon, especially in the up close, night battles that were typical of the fighting around Guadalcanal.

It was longer than 9 meters, could travel up to 20 miles at speeds up to 52 knots, and had a warhead in excess of 1000 lbs. This torpedo “out-sticked” (had a longer range) and outperformed the American Mark XV torpedo in all aspects. But, fortunately for the U.S. Navy, its range exceeded the range at which Japanese ships could detect their American prey. (1)

The reality was that most battles involving the use of Japanese torpedoes occurred at point blank ranges where the torpedo’s range advantage was completely negated.  This is not to downplay the torpedo’s lethality in any way.  It was a deadly weapon at any range.  However, it was not the long range threat that its capabilities indicated it could be.  If the Japanese had had effective long range radar, for example, to provide targeting data, the results could have been far worse for the US Navy than they were.

We see this phenomenon of weapons outranging sensors being played out repeatedly, today.

The Zumwalt’s 70-100 mile rocket propelled LRLAP munition (set aside the fact that it never performed to spec and cost too much) required fixed target coordinates and no one ever explained where those target coordinates were going to come from.

The Navy wants to equip its ships with the new, long range anti-ship cruise missile (LRASM) with a range two or three hundred miles but no way to consistently provide effective targeting.

And so on.

As with the Japanese torpedo, those weapons may still be effective at shorter ranges but that fails to take advantage of the full range capability of the weapons.

The lesson is obvious.  We need to devote as much or more effort to developing long range sensing as we do to developing long range weapons.  Further, the sensing needs to be something that is effective in high end combat, unlike the UAVs and P-8’s that the Navy seems to think will somehow provide targeting despite being easily detectable and non-survivable.

Sensors may not be as sexy as shiny new weapons but they’re far more important and the Navy has completely forgotten that.



___________________________

(1)USNI Proceedings, “Ten Seconds To Live Or Die”, Dec 2018,
https://www.usni.org/magazines/proceedings/2018-12/ten-seconds-live-or-die

64 comments:

  1. Actually the IJN had very good optical sensors. Their binoculars were far superior to that of the Allies. They also had highly trained visual watchstanders.

    If I recall rom "Japaneses Destroyer Captain" (Hara), the IJN wartime doctrine was to launch at 10-11 miles. Quite a bit longer than effective range of either the US MK-15 torpedoes or 5" guns.

    ReplyDelete
    Replies
    1. The problem is that in the night battles around Guadalcanal the darkness and the land masses prevented long range detection. Doctrine may have been 10-11 miles but reality was point blank range.

      Delete
    2. I agree that the Long Lance capability wasn't fully exploited. Although that maybe speaks to larger problems in IJN doctrine. Had they gotten the "decisive battle" they sought, the LL might've proved its worth.

      I don't believe the Navy's intends to use the P-8A for ISR&T in a contested environment. Or at least haven't read anything to that effect. It's primarily a long-range ASW aircraft. It's proven pretty effective at that - as did the P-3C for about 40+ years.

      Note: this is not to say that we don't need an ISR&T capability in contested environments. We do.

      Delete
    3. "I don't believe the Navy's intends to use the P-8A for ISR&T in a contested environment."

      The Navy has not explicitly stated that. What they've stated is that the mission of the P-8/Triton is Broad Area Maritime Surveillance (BAMS). What they've stated is that they want to implement distributed lethality which depends on sensor networks inside enemy territory. What they haven't stated is how they'll provide those sensors since there are no other broad area sensors. Logic clearly indicates that the Navy plans to use P-8/Triton as their main sensor in enemy territory.

      If that is not true then the distributed lethality concept has no hope of working (which it doesn't) and the Navy knows it. The only possible conclusion is that the Navy believes the P-8/Triton is capable of enemy airspace sensing which is just ludicrous.

      FYI, the initial P-8s are less effective than the P-3s but that's by design (technology lock) and ought to improve as the P-3's upgrades are eventually folded into the P-8s, assuming the Navy allocates the budget to do so.

      Delete
    4. "Had they gotten the "decisive battle" they sought, the LL might've proved its worth."

      This is a good example of doctrinal arrogance and failure to acknowledge that the enemy gets a vote. That the US Navy didn't obligingly line up for a decisive fleet battle under conditions favorable to the Japanese is something the Japanese should have anticipated and allowed for with alternate doctrine.

      This is also what the current US military is guilty of. We are absolutely arrogant about our assumptions and all of those assumptions assume complete cooperation by the enemy with our plans.

      Delete
    5. That's s very important point, the long Lance was a fantatic tool to fight the Last War.

      Had the Japanese and American gun ships sailed to a mutually convenient patch of ocean to exchange artillery, long lance would have been devestating.

      Unfortunately for Japan, America was unwilling to cooperate.

      Delete
    6. This reminds me I need to finish reading "Ship Killers". Great book that covers the design and intended doctrine for all major navies (UK, US, IJN and Germ) torpedoes.

      Delete
  2. Question; Would a small fast surface ship (Radar-picket corvette?) Be of value? Large radar and powerful comms, but a primarily self-defense weapons suite and a bunch of ECM/Chaff/flares.

    It will serve to both find targets for DDGs, and draw fire away from DDGs and larger ships. Let the Destroyer pack the offensive punch.

    I'm just spitballing ideas here.

    ReplyDelete
    Replies
    1. Basically its an LCS with better radar that is roughly doing a suicide mission.

      Delete
    2. If you were going to do that, might as well use a UAV drone like the Fire Scout or something.

      Delete
    3. A destroyer is going to have a hell of a time defending a target 50 miles in front, just from missile flight time.

      Delete
  3. The Long Lance was a great torpedo but it was actually a little more dangerous to the attacker. One hit on the deck tubes igniting all the compressed oxygen and explosives could easily sink the ship. Many captain jettisoned the torpedoes at the sign of an air attack.

    The Japanese were so entranced by it that many ships had reloaded systems. That limited deck space, increased weight, and increased the dangers of explosion. Once the battles shifted from close quarter night ones to carrier vs carrier the torpedoes were next to useless.

    ReplyDelete
  4. "The Zumwalt’s 70-100 mile rocket propelled LRLAP munition ... required fixed target coordinates and no one ever explained where those target coordinates were going to come from."

    In an amphibious assault, target coordinates would come from forward observers on the ground. I guess they could also come from UAVs too.

    ReplyDelete
    Replies
    1. "In an amphibious assault, target coordinates would come from forward observers on the ground."

      Observing 70-100 miles deep into enemy territory???? What kind of magical, superhuman observers are these guys?

      " I guess they could also come from UAVs too."

      Haven't we covered this enough times? No UAV is going to be able to leisurely cruise over a battlefield thick with all manner of sensors and weapons and survive. How many times do we have to go over this?

      If you have some data showing that a UAV can survive the concentrated observations and weapons of a peer military that is looking for UAVs then present it. If not, simple logic tells us that the lifespan of a UAV will be very, very short. Stop repeating a fallacy that is nothing more than wishful thinking.

      Delete
    2. It's true you can't fly a predator or a global hawk in contested airspace without losing it. In a peer war enemy light non-stealth UAVs are just convenient target drones. To add to this, Beijing has a credible but improbable chance of detecting even 5th gen stealth bombers and fighters and successfully scrambling fighters during any attempted ingress of China's industrial regions. What they lack is targeting quality returns from stealth aircraft.

      Competent 1st world IADS and A2/AD are not the end of aerial C4ISTAR. The B2 bomber has superb stealth in addition to a *wonderous* defensive suite including LPI radar, passive EO/IR, ESM, stealthy ECM, and a stealthy, hardened comm link with computers that can deconstruct signals to reveal the components returning from both target and jammer(s). The B2 can identify radar installations by their emissions, including warships. It can fly around them, even if they're warships or unaware aircraft. It can get rather close before being detected, and its comm link will work right up until nukes start detonating - which is good enough!

      The RQ-180 - if it exists - is the same platform in UAV form, sans large thermonuclear payload; obviously more bombers are required due to the latter flaw. However, it's probably also worth investing in even if there's no proof it can do the same C4ISTAR mission as the B2 presently. The battlefield looks like we will desperately need a cheaper stealth ISR craft in the future even if there aren't efficient returns at present.

      If the RQ-180 exists, it and the B2 are a formidable stealth ISR system. Their electronic counter-countermeasure capability and the physics of their antennas are not to be underestimated. For smaller, cheaper UAVs this is not to be assumed, as CNOps so rightly reminds us. They're our best (and a winner's) bet for PGM targeting data acquisition in a contested area, followed closely - if less desirably - by hundreds of JTACs with a high attrition rate. CNOps rightly states that LRLAP can only reach a short range in front of the JTACs - even if they're only 3 miles inland.

      Zumwalt could maintain both USNs standoff doctrine and NGFS with LRLAP against targets 0-30 miles inland. If the B2s are considered, LRLAP could provide deep inland strike without JTACs and marines dying by the thousands to get there... if only LRLAP weren't a massive failure. I'm not arguing that LRLAP would make Zumwalt an efficient or necessary platform, but rather than many of the core concepts of BVR engagement in a heavily contested area can be achieved by current platforms. Replace LRLAP with tomahawks or other BVR weapons that actually exist, and the strike capability of a SAG or CVBG with stealthy ISR support is very real, even in a contested environment.

      Delete
    3. "In Vietnam"

      The point I think CNOps is making and you seem to be missing is that a peer war with China - the scenario which should dictate most of our acquisitions and doctrine - will be far less amenable to flying any kind of aircraft, and the attrition rate for marines and JTACs in an amphibious assault would be *unacceptable* if NGFS requires them to be the primary ISR asset. I think stealth ISR craft can fill that niche, CNOps favors 16 inch gun area bombardment and counter-battery fire, and other proven WW2-era methods of accomplishing NGFS and other strike missions. We all agree that JTACs have a role, but you do not want them pushing into visual range of pristine enemy defenses, you need to deliver ordinance further in front of them.

      Delete
    4. A ground based forward observer is limited to a view of a hundred yards or so on out to perhaps a mile, depending on terrain. A forward observer can't penetrate 70-100 miles through peer enemy formations to direct fire support. He's limited to our side of the lines and a pretty short view across the line.

      Airborne observers can, theoretically, penetrate but not survivably.

      We need to do two things:

      1. Develop new forms and methods of survivable, deep penetration surveillance and targeting.

      2. Reinstitute heavy caliber, area naval bombardment which eliminates the need for precise targeting.

      Delete
    5. Luke, For its time, Vietnam had a pretty sophisticated air defense system and there was often the threat from small arms as well. While that influenced our training, tactics, and procedures, it didn't from flying. The underline notion seems to be that China is all knowing and all seeing and I just don't believe that.

      In support of the raid to release our hostages in Iran, the Air Force flew a light aircraft into Iran to scout locations for what became Desert One and Desert Two. The scouting included landing and evaluating the terrain and laying out landing aids for the rescue force. Besides, we spent 40+ years preparing to fight Russia and the Warsaw Pact. China poses a formidable threat, but they can be beat.

      Delete
    6. "Vietnam … Iran"

      What is your point? Is it that because we could fly over Vietnam and could fly into an uninhabited region of Iran that we can fly wherever we want in a high end, peer war with China? If so, that's ridiculous.

      I would also point out that we lost a couple thousand fixed wing aircraft in Vietnam. How much worse would it be against a peer opponent?

      I'm truly at a loss as to what your point is.

      Delete
    7. My point is that China is not the impenetrable fortress you make it out to be. Neither is China is all knowing or all seeing. If we intend to defeat China or any other enemy, we have to get into their backyard, whether that is on land or on the sea. And, we have the troops and resources to do just that.

      We spent 40+ years preparing for a peer war with Russia. And, we were prepared to lose thousands troops and aircraft in the opening phase of that war. Hopefully, we would be better prepared and fair better in a war against China. But, we can't be afraid to engage an enemy wherever they are.

      Yes, we lost a couple of thousand aircraft in Vietnam, but that didn't stop us from flying.

      Delete
    8. Why are you suggesting that we should be afraid to fight China? No one around here believes that.

      Delete
    9. "Why... Afraid to fight China?"

      We often refer to the USN's standoff doctrine, the do-everything-ship-death-spiral, difficulties employing other non-attritable assets, limited numbers of attritable platforms, and other issues in US budgeting and acquisitions. I see how it can come off as being "afraid" to fight China. We're certainly afraid (or at least deeply unwilling) to lose to China, and a lot of scenarios look like we're heading for just that with our current force structure... at least in a "limited" peer war, which many commentators here and elsewhere seem to prefer and/or think is most likely. I maintain that we have a distinct advantage if *we initiate* Total War and accept the military and industrial costs that it entails.

      What I was trying to get at in my comments is that if we do invest correctly in attritable platforms, proven conventional systems like battleships, and effective C4ISTAR platforms then we can also engage in limited war without the kind of risks/costs that US decision makers seem unwilling to accept.

      Delete
  5. Whilst it may appear that weapon ranges typically run ahead of sensors, there are classified programs that may provide the targeting information in the kill chain that we just don't have privy to. Some examples that come to mind are the NOSS satellite constellations; the RQ-180 program; the B-21 platform acting as sensor nodes; and even the F-35 providing the targeting information in the kill chain as part of the concept of "combat cloud".

    ReplyDelete
    Replies
    1. I would say you have problem if a not yet existing bomber of unknown functionality is important to making tactical decisions in the now. How many RQ-180s? A drone system that has never faced serious obstruction efforts.

      Delete
    2. " classified programs that may provide the targeting information"

      Sure. Maybe there are magical, invisible, all-seeing sensors that we know nothing about … … but I don't think so.

      Satellites can't provide targeting information. The time delay between sensor data collection, transmission, interpretation, and dissemination to a military unit far precludes useful targeting data. At best, they can provide a general sense of where some enemy unit was some time ago. People seem to think that satellites are direct linked to a missile on the ground. Nothing could be further from the truth.

      An F-35 provides a pinpoint area of sensing compared to an actual broad area sensor. Of course, the only effective broad area sensor is radar which pinpoints its own location. At best, an F-35 would provide targeting for an occasional target that it stumbles across and even that's unlikely because, again, the F-35 is not directly linked to a shooter.

      Magical, secret sensor platforms sounds like a TV show. We'll stick to what we know.

      Delete
    3. The idea of sensor shooter in modern warfare is not about individual platforms but a kill web i.e a networked kill chain. Such capabilities had been demonstrated using the TTNT architecture by the Navy's NIFC-CA system where the shooter was a DDG and the targeting data provided by an F-18. The TTNT architecture is a four-channel radio system using Link 16 but with enhanced throughput and wideband capabilities offering OTH engagements, integrating multiple systems,airborne and on the ground, into a single kill chain. In a denied environment, the F-35 acting as sensor will provide the link via MADL instead of Link 16. It is not some TV show as you alluded.

      Delete
    4. Do you have some point to make?

      Regarding the F-35 as a targeting platform, F-35s cannot, to the best of my knowledge talk targeting to ships or ground forces. No one has MADL other than the F-35. The Navy has conducted a few tests which were basically just unrealistic public relations exercises and they've talked about incorporating MADL in various platforms but have not done so, as yet. Link 16 could be used but that has a number of problems.

      Delete
    5. My point is simply that the sensor shooter asymmetry is partially addressed via networked solutions linking sensor and shooter.

      As to your opinion that the F-35 is a poor sensor targeting platform, I would be interested in some facts behind that opinion. Clearly MADL does present a problem when attempting to exchange LPI data with other platforms in the kill web. There are solutions in the form of using multi gateway nodes in BACN platforms. There are 8 in the inventory with 4 EQ-4B's and 4 E-11A's providing the links. Additionally, F-15's when carrying Talon Hate pods can communicate with the F-35's in LPI mode.

      Delete
    6. My first question is how does sensor platform aspect impact the LPI comm signal and how does the transmitter track the receiver? If the key to LPI communications is tightly directed radio signals, how does the F-35 maintain a signal while maneuvering? And does it require LOS to the receiver?

      If the F-35 (or any other platform) has to remain at a specific aspect and altitude, this greatly limits it's range. Signal LOS is a square root function meaning that the antenna has to be 4x as high to have 2x the range. So a ~100nm LOS (either radio or radar) would require an altitude of ~6,000 feet.

      And it seems to me that either the F-35's sensors are highly automated or we're seriously over-tasking the pilot by having them act as a sensor operator as well.

      Delete
    7. All good questions, none of which I know the answer to. Maybe someone else can jump in?

      Delete
  6. I thought theoretically the F=35 was supposed to fill the role of long range sensor platform.

    ReplyDelete
    Replies
    1. The F-35 is not a broad area surveillance platform. It offers a soda straw view, by comparison. Under certain circumstances that can be helpful but it is not an area surveillance asset.

      Further, the most effective area sensors are large powerful radars. The F-35 would be using small, limited field of view, limited range non-radar sensors - hence, the soda straw view.

      Delete
    2. I agree I just failed to catch post. I not saying it would/will work. I am just saying its a talking point the pentagon is selling. I might note it is also one of increasing trend of saying a one pilot fighter can do all kinds of jobs at the same time with perfectly equal assumed ability.

      Delete
    3. It has to use a soda straw view otherwise the whole ideal is pointless. It probably is. I'm not sure how an F-35 or 2 or 3 are going to loiter about with only passive sensors and not be as equally vulnerable to the same and than stealthy communicate how exactly.

      Delete
  7. The problem ceases toe xist if sensor platform and munition launching platform are separate, yet have a a satisfactory communication link.

    Moreover, even sporadic effective observation may have a great effect; the risk of being detected and engaged may be unacceptable to the opposing force.

    An example for this would be if long range SAMs were used to deny air lift - this could work even if they have eyes on the airspace above the airport only 10% of the time (irregularly).

    ReplyDelete
    Replies
    1. This might be true for target acquisition, but target tracking generally requires a continuous sensor input. Airplanes move fast and you'll waste a lot of rocket fuel if you aren't vectoring the munition constantly. This problem gets worse the further the munition has to travel to get to the target area and/or the longer it takes.

      And the more complex (and expensive) a system is, the less tactical it becomes and therefore the more susceptible to chain of command delays and approvals. If you are shooting off a million $$$ missile, you aren't going to want to waste it on a liason helicopter taking off. Who determines if a target has sufficient value to warrant expenditure of a precious 200 NM SAM?

      Delete
  8. Submarines and Satellites can provide excellent targeting data. Coast watchers and tiny fishing boats too. Sigint too, unless one thinks Carrier groups will operate in radio silence, to include aircraft. Do they even practice this? And the new Ford will send out a massive electromagnetic signal with each launch.

    ReplyDelete
    Replies
    1. You're confusing targeting data with general situational awareness. A missile launch requires precise target coordinates. General situational awareness (ship movements, for example) is not launch quality data.

      SigInt is, again, awareness but not, generally, launch quality targeting data although, with enough cross bearings one might be able to generate launch coordinates.

      Delete
    2. If you believe what Raytheon and Kongsberg say about the NSM and its follow up the JSM and I reading it correctly... You fcould ire in the general direction and supposedly the missile will self target. But of course it could target a lot of things and you might need to fire a lot.

      Delete
    3. Yes, if you have pretty good idea where a target is, you can semi-blind fire and hope to get lucky. However, if the missile doesn't find anything it's a million dollar (or whatever the particular missile cost that you used) waste and now your inventory has one less of a hard to replace weapon. Commanders are going to want a pretty solid idea of a target's location before flinging million dollar missiles around. And, of course, there's the whole friendly/neutral target issue that you mention.

      The ocean is immense. The missile's field of view is miniscule.

      Delete
    4. "If you believe what Raytheon and Kongsberg say"

      If we believed what was claimed about the Ford …

      If we believed what was claimed about the LCS …

      If we believed what was claimed about the Zumwalt/LRLAP …

      If we believed what was claimed about the F-35 …

      And so on.

      Delete
    5. @Kath: this capability is nothing new, you could do the same thing with TASM and Harpoon back in the 80s. Blindfire a missile and it goes after the first contact to enter it's seeker cone - if it even sees anything.

      Delete
    6. Well, technically it's a "bearing only launch". You know the basic direction to fire, but not the specific range. This is predicated on the munition having sufficient sensor capability and speed to localize and adjust to the target's location. It's not too bad for firing at ships as you can fire a sea skimmer in the general direction of the target fleet and it only has to "look ahead" as opposed to trying to pick out a target in the midst of ground clutter.

      Delete
  9. Couldn't a sub provide coordinates for an anti-ship missile?

    ReplyDelete
    Replies
    1. Or an E-2 at 25,000 ft looking oth?

      Delete
    2. A sub could although it would have to broadcast to do so and subs don't like to do that.

      Delete
    3. An E-2 can certainly provide targeting data but it would likely be shot down long before it could do so and no one is going to risk a valuable E-2 trying to conduct deep penetration search.

      Delete
    4. Shot down by what? Not every enemy surface combatant is going to have air cover. Why couldn't an E-2 safely observe a Chinese SAG from 150 miles away? whats the max range of Chinese shipboard SAM's?

      Delete
    5. "Shot down by what?"

      We recently did a post on the new Chinese Very Long Range Air to Air Missiles (VLRAAM) with ranges of hundreds of miles.

      E-2s are going to be one of the highest priority targets for the Chinese. Given that combat will occur within the Chinese controlled air space (E/S China Seas and first island chain), there will be plenty of air cover for all Chinese ships.

      Chinese destroyers carry the HHQ-9 family of SAMs with a range of 125+ miles.

      There is no evidence that an E-2 could detect a stealth destroyer at 150 miles. It might, it might not. We have no idea.

      So, yes, one way or another, the life expectancy of an E-2 outside the protective sphere of the carrier battle group would be very short.

      Delete
    6. The benefit of the E-2 as a sensor platform is it's height. This gives it a much larger radar horizon. But it also means it's a giant EMR lightbulb above the fleet.
      So you station the E-2 on the threat axis and a short distance away from the fleet. You have your fighters there as well as that's where you expect the enemy to strike from. The purpose is to give the fighters as much lead time to an inbound raid as possible to vector for interception.

      Delete
  10. 'Sensor and AI'. Today, commercial autonomous cars are common place, and they're about equivalent of average human 'eye/brain' in see-think-act. I can see in very near future (if not already) of autonomous 'long lance' or other over the horizon munition.

    ReplyDelete
    Replies
    1. To the best of my knowledge, current autonomous technology cannot distinguish between stationary objects in the road versus those near the road. For example, between a broken down semi-truck in the lane ahead and a road sign over the lane.

      In addition. they all rely upon very precise and detailed mapping data which is compared with GPS data to be able to navigate.

      The systems also can only react. They can't, for example, identify a car driving aggressively and take steps to allow extra space in front in anticipation of the driver cutting in front of you to pass.

      Humans are very good at pattern recognition. In some cases too good as we try to find meaning in random events. For example, AI's don't have lucky shirts.

      Delete
    2. Youtube has videos of self-driving examples of realistic road conditions. As a human driver, I also need information to direct where I'm going, either from memory, a map, or GPS- all of which can be replicated by robot. As for 'exceptional events', as a human driver, experience helps (e.g. a newbie driver vs. someone with decades of experience) which also can be learned by robots.

      I think what I want to say is: instead of 'long lance (must have firing resolution)', how about 'silicon Kamikaze (general target location will suffice)', I think we are a lot closer to 2020/30 to 1941/45.

      Delete
    3. " I can see in very near future (if not already) of autonomous 'long lance' or other over the horizon munition."

      I'm not really sure what you're looking for. We've had homing torpedoes for many years. Russia has specialized wake homing torps. Everyone's submarine torpedoes have "over the horizon" range. So, I'm not really sure what else you're asking for.

      Even if we had the functional equivalent of a human brain inside a torpedo, I'm not sure how that would improve matters. The likelihood that anyone is going to blind fire a torp 'over the horizon' just to see what happens is slim - it's expensive and wastes one of a very small inventory of ship weapons. You still need a pretty good idea of where a target is and, if you have that, we already have all the 'intelligence' we need in a torp.

      So, what are you looking for?

      Delete
    4. I'm using the wrong example, 'long lance' that is (unless it's a smart mine/torp lay in waiting & listening). The gist of your article (how I understood) is that non-reusable over-the-horizon munition (under or above surface) needs precise firing resolution. Take it a step further, a 'silicon Kamikaze' only needs to know target's whereabout. Therefore, the merger of 'sensor & AI' is probably closer to where we are today than, say, long distance munition needing precise firing resolution of yesterday year.

      Delete
    5. You're getting more and more lost. Unless you're willing to waste expensive and rare weapons on a 'hope' shot, you still need pretty precise pre-launch targeting. There's no getting around that. Plus, launching runs the risk of broadcasting your own position.

      The seeker head of a torpedo has only a very limited field of view. Again, you need a pretty good idea of where the target is before launching. While a Kamikaze pilot flying at 20,000 ft has a moderately wide field of view, a torpedo has a very restricted field of view. The tiny sonar (or other seeker) in a torpedo with its tiny power source has very little power and thus the sonar signal is weak and short ranged. No amount of AI can compensate for a very small field of view.

      So, still not seeing where you're trying to go with this.

      Delete
    6. During WW2 Okinawa campaign, Kamikazes from Japan flown over 700 km (couple hours flying time) by pilots who had <50 hr logged time, had never seen US fleet operation before, and they (or escorts) managed to find the fleet based on just general bearing and distance info (gathered from who knows what against total US air/naval superiority.)

      Consider today's peer opposition's 24/7 all weather/global surveillance capability, I don't think anything can be hidden (above surface), after all, trainings can't be hidden. And we've seen couple years ago, an AI can best AF instructor on the simulator, thus it's not hard to extrapolate what AI can do in other fields. Now, combine that with AESA and optical systems, there you have 'silicon Kamikaze' on future high subsonic or super/hypersonic munition targeting ships that are not much faster or maneuverable, and even less armored, than their predecessors.

      As for 'long lance', it might already be in the form of 'autonomous subs' with details unknown to the public.

      Delete
    7. Remember the Kamikazes hit more DDs and DEs than CVs. Many escorts were hit by multiple Kamikazes.

      So how does the AI distinguish the correct target out of a task force? How quickly will we be able to update AI against new deception techniques? How do you test such an AI under near-combat conditions? And how do you protect the technical details of the AI since a potential enemy would have a huge incentive to obtain them?

      Delete
    8. Jay, Pattern/object Recognition (e.g. China's civilian surveillance system where AI can real time ID individuals based on facial features and movement gait, that's for at least tens of millions of city populations) shouldn't be an issue; there are only tens of thousands of ships (both civilian and military) in the world. The leg work is in collecting the database, which all sides are doing now, not just on the individual surface ship, but also the operation movement (all sides have to train and have exercises, therefore be observed).

      On 'disrupting' AI..well, the traditional kill-chain starts from, say, a SAT (or AWACs) down linking to shooter and then to the arrow, which has to be fed and directed until the last 40-50 miles (or within the horizon distance) before self-homing kicks in. The idea is to cut the chain before the last 40-50 miles. Well, back to ww2-Kamikazes, I'm fairly sure the USN had tried hard to keep the fleet whereabout from the Japanese, still the Kamikazes found the fleet. Therefore, today the AI, in each flying munition, diffuses the kill-chain and localizes the command within itself and the swarm perhaps (see drone swarm on youtube).

      As for 'how to test AI under near-combat condition', well, the same way you train a flesh & blood pilot: how do you train them under near-combat condition? During real pilot training, all information is recorded, evaluated, and improved upon. We(human) filed away in our brain, AI learns the same way. As for 'protecting the AI details', different algorithms (as far as I know) are six&half dozen differences, the main draw is the database (or, in our human parlance, the experience), the longer we see each other train/moves, the bigger the database collected.

      CNO brought up the expense issue of AI-missile. Well, a $30M F-18 or $100M F-35 (and the $15B carrier which must tag along with them) is worth between 6-20 AI-missile (say, $5M a pop). The question then becomes: can a F-18 or F-35 survived 6-20 combat sorties (against near peer defense) to get its money worth? I don't know the answer.

      Delete
    9. I still have no idea what, if anything, you're calling for. You're describing, inaccurately, the Kamikaze experience but not tying it to anything - at least, nothing we don't already have.

      You know, don't you, that the American fleets at Iwo Jima and Okinawa (presumably the Kamikaze attacks you're referring to) were in a known, essentially fixed location for the amphibious assaults. They had to be. It's the nature of an amphibious assault. You have to be located right next to the place you're assaulting. Thus, there was no mystery about location and no question of the Kamikaze pilots finding the fleet. The only question was whether the pilots would survive long enough to find a target ship.

      We already have semi-autonomous missiles that can detect a target. Of course, those missiles are fairly stupid and relatively easily decoyed or otherwise fooled/defeated. Optical pattern recognition, for example, is easily thwarted by simple obscurants (smoke, both simple and multi-spectral). Radar is harder to fool but still susceptible to decoys. ECM is highly effective.

      So, I'll ask again, what are you asking for? What point are you trying to make?

      Delete
    10. What I'm trying to say is: in the near future, decision making will shift towards where the munition at, not where the shooter was, much like a pilot flying his mission. The only difference being: air plane can come back, an AI-missile is like a Kamikaze of old, except AI will be a much 'experienced' pilot. Come to think of it, the 6-gen will be a reusable AI-platform, so between what we had before (i.e. semi-autonomous munition) and 6-gen AI-plane, there may very well be a slot for non-reusable AI-platform (i.e. AI based munition.)

      Btw, I foresee a future where naval battles will be fought (or at least capable of) in storming pitch darkness, where eyeballs/brain, biological or silicon, will fight what computers/sensors tells one to.

      As for naval engagement decision or format, it all depends on the strategic objective of the combatants, still, targets have to be found and engaged, and I do believe AI will play increasing role in that.

      Delete
    11. Either you're not conveying your idea or you're unaware of what already exists. For example, Aegis, in full auto, makes all the decisions about target priority, weapon allocation, launch timing, etc. - which seems to be exactly what you're calling for. If so, it already exists.

      We already have missiles that fly waypoints and then search for targets, all autonomously.

      Well, enough of this ...

      Delete
  11. One general comment I'll make is the confusion regarding strategic versus tactical sensors. Satellites are strategic sensors and their data goes to strategic level commanders and decision makers.

    Generally, the more "hops" there are between the sensor and the use of that data, the more strategic the sensor is. Satellite data has to be interpreted and analyzed before it is useful. Likewise for most SIGINT. The other factor is how rare or expensive a sensor platform is. Satellites are expensive, rare and can't be quickly replaced. Their data links are also expensive, immobile and large. Therefore their priority is high value data for strategic decision makers, not for target individual ASH missiles.

    One common complaint about the US military is that intelligence (including sensor data) is kept too high up the chain of command and too stovepiped from other consumers. As such, command decisions keep being moved up the chain of command.

    ReplyDelete

Comments will be moderated for posts older than 30 days in order to reduce spam.