Monday, November 12, 2018

Sensors - What Good Are They?

We’ve previously noted that the US military has gone all in on sensor networks and unmanned vehicles as the basis for its Third Offset Strategy which is intended to provide the US with a military advantage over its enemies.  We’ve also noted the Navy’s commitment to distributed lethality which also depends on regional sensor networks to provide targeting to roving ships with a few anti-ship missiles.  Further, the entire basis of the F-35’s hoped for ‘superiority’ is sensor fusion and situational awareness (it’s sure not air combat maneuvering!).

Unfortunately, we’ve also noted that the entire concept of sensor and data networks is inherently flawed.  Non-enemy induced network crashes, sensor failures, UAV communications link failures, and GPS failures are commonplace.  Add to that wartime cyber attacks, jamming, GPS disruption, electronic countermeasures, electronic spoofing, false signal injection, etc. and the prospects for successful sensor and data networks is dismal – and yet we’re betting everything on exactly this.
Are we being overly pessimistic?  Well, consider,

• The Russians have been giving an object lesson in electronic warfare in the Ukraine and Syria.  US commanders have acknowledged that the Russians have disrupted and ‘disabled’ our dedicated electronic warfare EC-130 aircraft over Syria.

• Iran is believed to have disabled and captured US UAVs.

• US Navy unmanned underwater vehicles routinely wander off due to communication’s loss, never to be seen again.

• Despite having inertial navigation systems and GPS, the Aegis cruiser Port Royal managed to run aground in well known, well charted, home waters.

• Despite GPS and extensive regional sensors, two US Navy riverine boats became lost and wandered into Iranian waters and were captured.

• The Vincennes shot down an airliner despite having continuous, unhindered radar contact.

• Despite the most advanced naval sensors in the world, the Navy has been unable to determine whether any of three separate attacks on a Burke class destroyer off Yemen actually occurred.

• Despite the world’s most advanced radar, sonar, and electro-optical sensors two Burke class destroyers managed to collide with large, slow, non-stealthy commercial ships.

The examples are nearly endless.

Now, as has been recently and widely reported, the Norwegian Nansen class frigate Helge Ingstad (F313) has collided with an oil tanker and been beached to avoid sinking.  The ship has essentially capsized and is laying on its side on the beach.

The Nansen class frigate possesses a multitude of advanced sensors of various types including,

• SPY-1F 3D multi-mode radar
• Reutech RSR 210N air/sea surveillance radar
• Sagem Vigy 20 Electro-optical
• MRS 2000 hull mounted sonar
• Condor CS-3701 ESM/ECM

Despite this impressive array of sensors which should have provided unparalleled situational awareness, the ship managed to collide with a tanker. 

On a related note, yet another modern ‘warship’ has been nearly sunk by a single ‘hit’ – not exactly a tribute to modern warship design, is it?  But, I digress …

The empirical evidence is overwhelming.  Our vaunted sensors and networks do not work at anywhere near the claimed levels.

Our sensor and network systems are simply not reliable.

They don’t work.

And yet, we’re betting our military future on them working flawlessly and doing so in the face of a vast array of countermeasures.

The reasons for failure are many and varied and not all of the failures are due purely to the sensors and networks.

• Comm. links fail (UAVs being lost)

• Networks spontaneously fail (we’ve all experienced this at work or in the military)

• Maintenance shortages cause degraded sensors (the Aegis system being a prime example)

• Human action based on sensor data is inherently flawed (Vincennes)

The overall conclusion is that we can’t count on sensor and data networks and we can’t count on having situation awareness – and yet that’s exactly what we’re betting our military future on.  It’s also worth noting that all of the examples of sensor and data network failings are peacetime examples when everything should work perfectly.  How much worse will our situational awareness be during war?

Now, having said all that, I’m certainly not suggesting that we should abandon sensors and networks.  What I’m saying is that we should acknowledge the inherent limitations and tendencies to fail and not bet our military future on them.  Instead, we should use them as adjuncts to basic technologies (binoculars or sextant, for example) and common sense (post lookouts!).  We should train to function without sensors, to any great extent, and then we’ll be pleasantly surprised when they do, occasionally, work.  We have to break our mindset of dependence on sensors and networks and learn to stick out heads out the porthole and look and reason for ourselves.

We also need to recognize that data, alone, is useless in war.  We need firepower to destroy whatever we see.  Failing that, we’ll have the most perfect awareness in history of the enemy that kills us using low tech, indiscriminate, area bombardments.  For all its impressive development of electronic warfare capabilities, the Russians have not neglected to also develop impressive families of armored vehicles, advanced cluster munitions, treaty busting cruise missiles, very long range air-to-air missiles, advanced torpedoes, etc.  They understand that, ultimately, firepower wins wars. 

25 comments:

  1. I feel I should note that you can have the best sensors in the world, but if you turn them off and don't use them, and when your crew is complacent even when the tanker is frantically warning of the imminent collision, all those fancy sensors can't do much to help you.

    Radar and AIS can't warn you of an approaching ship if both are turned off!

    ReplyDelete
  2. I chuckle at this because I have friends who laugh at my anti sensor rants. In most cases it's because I long for the days when a sensor just told me 'Low coolant' and not 'I detect low coolant and I'm shutting the car down without consulting you...' only to find out the sensor itself went bad.

    We've built these very complex sensor grids, and I suspect it also extends to the Navy's desire to be minimally manned (we don't need a watch stander to look at that gauge! We've replaced him with a sensor that reports to one person looking at 20 different gauges!'

    Often times it seems we expect the sensors to do the work, too. It's no longer a guy looking at a scope; it's the guy looking at five scopes and hoping one of them alerts him if something bad is detected.

    I actually think that they might cost us some situational awareness as we become, for lack of a better word, sensor blind. We are waiting for the sensors to tell us something that we could look out the window and see. It also leaves us screwed in terms of DC. If 2 out of 5 sensors tell the guy there is a major issue he can't just start working on the issues, he has to triage it. And then maybe abandon the other 3 sensors' data while he works on fixing the issues reported by sensors 1 and 2.

    Sensors are a good thing. But they are a tool.

    ReplyDelete
  3. I realize in the above I'm speaking about ships status sensors mainly, but I also meant to have it apply to the situational awareness sensors of the ships; radar, etc. Yes, good to have it. A necessity even. But it's just a tool and shouldn't be the only thing in the shed. As you stated; watch standers should still be there.

    ReplyDelete
  4. Distributed lethality reminds me of the 8-inch guns mounted on the casemates of interwar carriers: if it comes down to them, then you've already lost.

    ReplyDelete
    Replies
    1. A key part of Distributed Lethality is returning AShM capability to the surface fleet; for the longest time only 1/3rd of the Burkes has AShM canisters; the Flight II Burkes gave up their missiles for the helo hangar. Returning AShMs to the DDGs mean you complicate targeting for the opposition: now evernnship in the CAG is an ASuW threat to the enemy SAG, not just the fighters.

      Delete
    2. "Returning AShMs to the DDGs mean you complicate targeting for the opposition"

      Absolute nonsense. Whether or not a capital ship has anti-ship missiles does not determine whether the enemy will strike at it. The enemy is going to attempt to sink every ship, regardless of its exact weapons load.

      In modern naval combat, surface warfare between opposing ships is going to be the least common type of combat for a host of reasons. A DDG's (Burke) Aegis load is far more of a threat to the enemy than the several extra anti-ship missiles a Burke might carry. Thus, the enemy is going to consider the Burkes high priority targets regardless.

      Delete
    3. Hence the 8-inch guns on carriers analogy, Goose.

      Delete
    4. @Atlantic Turtle: I mean, you *can* see where the USN is coming from; the theory is you can do multiple vector attacks, but it's also that USN risk mitigation thing. There's an argument that the USN builds multirole ships instead of specialist ships as a form of risk mitigation.

      I'd liken AShMs on a DDG to torpedoes on an Atlanta, myself. *shrug*

      Delete
    5. " the USN builds multirole ships instead of specialist ships as a form of risk mitigation."

      Absolute nonsense. You can't mitigate risk by concentrating functions on one platform. That increases the risk by having fewer ships that are more expensive and the loss of any one is a bigger disaster. You mitigate risk by dispersal and increasing numbers. A Burke, under anti-ship missile attack, is at no less risk because it can also do ASW.

      Delete
    6. "8-inch guns mounted on the casemates of interwar carriers: if it comes down to them, then you've already lost."

      Quite right. It would be a very unusual, one-in-a-million type scenario in which a few guns on a carrier would be beneficial.

      This gets back to the need to design warships for the reasonable threat scenario not the one-in-a-million. The problem with designing to the one-in-a-million is that it is prohibitively expensive because you've got to include every conceivable capability because you just never know what you'll encounter. That's unaffordable and leads to too few ships - exactly the situation the Navy has placed itself in today.

      Good comment.

      Delete
  5. I would like to propose that given the value and purpose of the DDs, which is primarily air defense, a much smaller surface warfare ship of about 2000 tons with 20 VLS could be used to locate and attack an enemy fleet when the electronic sensors fail. Target data of enemy ships could be sent to the CVNs and DDs using open air radio.

    Also, the small scout/attack ships would have 20 heavy anti ship missiles each, to use offensively when contact is made. These ships would not have to extend beyond the air coverage provided by the CVN unless a specific mission called for independent action.

    The use of VLS would also allow land attack missiles to be carried instead of anti ship missiles freeing the DDs from this mission, which will allow the 'Burkes to carry more AAW missiles.

    This is a force multiplier beyond the simple addition of VLS numbers. It also takes into account the worst case scenario and gives the Navy an offensive capability that is not dependent upon the CVN or electronic networks.

    ReplyDelete
    Replies
    1. "a much smaller surface warfare ship of about 2000 tons with 20 VLS could be used to locate and attack an enemy fleet when the electronic sensors fail."

      You go on to say that these small ships would not leave the air coverage of the carrier. If that's the case, the ships are not needed because the carrier's aircraft can attack surface ships much more effectively. Each aircraft can fly hundreds of miles and launch anti-ship missiles that can fly another hundred-plus miles. No need to risk ships.

      You mention a worse case scenario. If the carrier can't launch aircraft, that would, indeed, be a worse case scenario and the carrier and its ships shouldn't be anywhere near a combat zone. If the carrier's launch capability has been somehow entirely negated, the carrier is probably sunk and its escorts along with it.

      Use of a dedicated land attack VLS ship has some merit but we already have those - they're Virginia class subs and SSGNs. Each Virginia can carry up to 40 missiles, depending on exact version of the sub. Same with the Los Angeles class although different missile numbers. We have dozens of subs - no need to build more ships to duplicate what we already have.

      Delete
    2. @KH: We've talked about this before. If you didn't have the land attack requirement, 64 cells is an adequete number for an AAW DDG, especially when that AAW DDG is working with other warships. You don't *really* need 96 cells for a pure AAW ship. Nice to have, but not really needed.

      Delete
  6. Hi ComNavOps, great post as always. Overall I agree that our degree of dependence on sensors and especially networked sensors is bound to leave us unprepared for a peer conflict where we expect these sensors and networks to come under electronic and cyber attack. However, as you also note, we shouldn't (and really can't) completely do away with advanced sensors. Just as sensors are useless without firepower, firepower is useless without sensors.

    Sure, the good old Mk1 Eyeball can direct non-OTH weapons against slow moving targets - like naval guns targeting an enemy ship - but it can't guide long range missiles or any kind of air to air missile, the requisite targeting data just isn't gathered by that system. Clearly, no number of sextants and/or binoculars can replace the capabilities provided by advanced sensors, and a ship limited to line-of-sight weapons would be hard pressed to stand up to any competent adversary with functional OTH weapons and sensors.

    Of course you briefly acknowledged that we shouldn't do away with advanced sensors, but rather be prepared for if they don't work. However, I don't see how any kind of preparation can make a sensor-poor (or sensor-failing) ship competitive with a modern warship with functioning sensors. It would be great if we could spend some time and money to make our sensors more reliable and our crews more competent at making appropriate use of sensor data, but I recognize that that's not something which is guaranteed to improve with funding/time. On the other hand, is there anything you have in mind that could keep ships with malfunctioning sensors relevant in an OTH missile exchange or other typical scenarios expected in modern naval combat?

    ReplyDelete
    Replies
    1. "is there anything you have in mind that could keep ships with malfunctioning sensors relevant in an OTH missile exchange"

      My compliments to you. You read and understood the post and you asked the right question. Here are some general answers:

      -We need to continue to develop sensors but we need to recognize their inherent limitations and vulnerabilities. Thus, we not only need to develop sensors but we need to work on making them and, more importantly, their comm links robust in the face of countermeasures. It's the latter aspect that we've largely ignored, so far. This likely means that we'll have to simplify our sensors and data streams. Instead of having a sensor transmit/receive that beautiful, full color, exquisite screen full of data that looks so magnificent in brochures but requires vast bandwidth - that we won't have in combat - we should be developing the MINIMUM data transmission in the most basic form that requires the least bandwidth and transmission time (there's your robustness!).

      -We need to develop sensor systems that are as simple as possible and have as few links (each link is a potential weakness and failure point) as possible. Instead of using an F-35 to find a target and then relay that data to a UAV that will send it to a central command ship/point for incorporation into a master, regional level situational awareness assessment and then transmit that large, mostly irrelevant picture to the shooter ship, we should be cutting out all links except the direct one. In this example, each ship should have its own cheap, expendable UAV surveillance drones that reports directly back to the shooter ship that launched it. Keep the comm and network chain as simple and robust as possible.

      -We need to work on EMCON operations. Radiate nothing so the enemy has to radiate to find us, thus revealing their own position. Currently, we have no idea how to operate silently. We have neither the equipment (much of our equipment CAN'T be made EMCON silent!) nor tactical training to operate silently. We had this in the Cold War. We could operate an entire carrier group and launch a complete strike without a single transmission - not anymore!

      -We need to emphasize passive detection. Passive detection has "unlimited" range and requires only triangulation to be effective (sometimes, not even that). We need to become comfortable operating passively. We need to install 360 degree EO sensors to functionally replace radar (think an aircraft's IRST but 360 deg on a ship - these already exist for ships but we haven't pursued them).

      And so on.

      Did that give you an idea of the direction we need to go?

      Delete
    2. (1 of 2)

      "My compliments to you."

      Thanks! I'm not a professional (I work in biology research), but I've been using a lot of my time to study everything military from the physics of tank armor interacting with KE-penetrators to grand strategy and nuclear counterforce for almost a decade. I must admit my study of SOPs and daily military life is limited because those topics are difficult to study from outside the military.

      "Currently, we have no idea how to operate silently. We have neither the equipment (much of our equipment CAN'T be made EMCON silent!) *nor tactical training* [emphasis added] to operate silently."

      This is a recurring theme. From your blog and a few questionable but reputable news articles (about ICBM launch crews using LSD, officers falling asleep at their posts, etc.) I am getting the sense that less than half of radar operators, missile launch officers (nuclear and conventional), helmsmen, navigators, or communications officers know how to do their jobs effectively enough for the Warship System (ship + crew) to get much of any benefit from an incrementally better sensor system, compounding the insanity of spending large amounts of money on and designing war plans around sensor networks that are susceptible to ECM at least, and crash at the start of battle at worst. It might not be so insane if we had the human and/or AI information-processing throughput necessary to actually use *all* (or most) of the multi-platform multi-band sensors meant to enable Distributed Lethality, but we don't yet.

      I was not aware that a lot of equipment couldn't be made EMCON silent. Obviously reactors and turbines always make noise, good passive sonar can pick up loud ships OTH, and acoustic stealth requires a ship to be designed for it from the start. EM stealth on the other hand I thought was a simple matter of shielding anything that doesn't need to "talk" to the outside, like your intra-vessel comms and computers. Things that do "talk" to the outside like Radars, EO, EM illuminators, and comm antennas mostly have significant RCS, but they aren't emitting when they're off. What components have detectable OTH EM emissions? I don't think really low-end EM emissions would be relevant if you're in visual range of an enemy ship, and only somewhat relevant for aircraft. The same thing goes for non-EM emissions; if it's only detectable LoS then it's probably not going to be what gives you away, that'd be your IR signature and/or your superstructure's paint job. I play CMANO and I don't think it models unintended emissions from EMCON-silent ships; it'd be interesting to see if I can add it in.

      You've emphasized Effective Training several times before, and I wonder if this could be one of the keys to improving the Warship System's effective sensor capability. About a third of the "sensor failure" examples in this thread and in the past few months were directly attributed to user error, and I suspect another third is indirectly effected by poor readiness, and could be prevented with Effective Training. However,

      "We need to install 360 degree [Electro-Optical sensors to functionally replace radar (think an aircraft's IRST)."

      Yes, combined with Effective Training, this would exceptionally enhance Line-of-Sight (LoS) situational awareness. It's like having a rear-view camera in your car, except it's everywhere... on your warship. If spaced only 10m apart, you would probably have triple-redundancy, and these things are tiny, low-power warship modules (sorry, M word), so it is not unreasonable to build in this way.

      (Cont...)

      Delete
    3. (2 of 2)


      "each ship should have its own cheap, expendable UAV surveillance drones that reports directly back to the shooter ship that launched it. Keep the comm and network chain as simple and robust as possible."

      +

      "Passive detection has "unlimited" range and requires only triangulation to be effective (sometimes, not even that)"

      UAVs are essential because EO, RADAR, and most other electromagnetic signals are LoS-limited - radar extends just beyond the LoS horizon, because physics. A *Really Tall* ship mast at 300 feet has a visual horizon against a 50 ft altitude sea-skimming missile of 28 miles, and a radar horizon against said target of 32 miles. A 50,000 ft high UAV has a visual horizon of 280 miles against said target, and a Radar horizon of 324 miles. They can travel that distance away and see that distance further, so if you have enough platforms to pursue that CONOP then you get a nominal EO range of 560 miles (assuming an EO comm link limits your range) and a Radar range of 648 miles. Building cheap, reusable, Medium UAVs that can exceed 50,000 ft with such an ISR suite is Very Hard. If you could do it, those ranges increase. Even a low-end UAV kept a few miles from your ship (so you don't give yourself away - The Horizon Works Both Ways) at 30,000 ft gives a massive improvement in detection range.

      It seems electro-optical sensors have similar range performance to Radars. What I believe they lack is the range-finding ability, which depends on active pings that we precisely time against a low background, whereas EO uses passive intercept of nearly-constant IR and visual light emissions that we don't have precise timing for, because the enemy made them, not us. What would be excellent is a UAV with active and passive Radar as well as IRST or even a full EO suite and a large, secure comm link. Ultimately aircraft are ideal sensor platforms because of the Horizon Problem.

      Along those lines, I think we should seriously consider using cheap UAVs to make active radar pings to cue semi-active radars (passive detecting third-party active) on our ships. We can get the benefits of active radar while our most expensive platforms are EMCON silent.

      "we should be developing the MINIMUM data transmission in the most basic form that requires the least bandwidth and transmission time (there's your robustness!)"

      +

      "Instead of [sending sensor data] to a UAV that will send it to a central command ship/point for incorporation into a master, regional level situational awareness assessment"

      What is the minimum data that the ship needs? Probably just the position, velocity, and classification (or an estimate based on sensors and a tag for the confidence level and sensor data types) of all of the relevant targets. But it takes a C4ISR hub to turn raw sensor data into a meaningful enough picture of the battle space to discern what targeting data is worth sending to which launch platform. If we don't cram a C4ISR hub (at least a small one) into our UAV, it needs to send a lot more of its sensor data to the ship so it can do that work - which it can. I don't have a strong (or well-informed) opinion about which is better yet.

      Delete
    4. My Google profile is weird, this is the same account as Luke C.

      Delete
    5. An alternative which have advocated before is a towed/tethered electric drone attached to ship, the thin cable supplies the drone with power and high bandwidth data communication. The French ECA group have one that flies at 100-150m / 300-500 ft., giving line of sight of ~50+nm. As everything it's a trade off compared to cheap less capable expendable UAVs which have more range, whereas tethered done 24 hours surveillance? Perhaps both :)

      DARPA experimenting with a towed para-sail.“TALONS’ surface-track radar extended its range by 500 percent—six times—compared to its range at sea level. Its electro-optical/infrared scanner doubled its observed discrimination range. The TALONS team plugged in a commercial handheld omnidirectional radio; that radio’s range more than tripled.”

      Delete
    6. Towed gliders can give continuous surveillance at a couple hundred meters, and should be comparable to self-propelled UAVs in price, if not cheaper. However, 50 km LoS is a marginal improvement compared to 300+ mile LoS at 50,000 feet. It's worth putting 2-3 medium-size UAVs to provide better continuous coverage. A physical data link seems appealing, but that level of bandwidth is overkill if CNO's suggestions about minimum bandwidth hardened data links are followed.

      Delete
    7. "I was not aware that a lot of equipment couldn't be made EMCON silent."

      Once upon a time, every piece of electrical equipment on a ship had to be EMCON capable. That has fallen by the wayside. A major example is the Ford's EMALS catapult system. It uses hugely powerful electric motors that give off omni-directional, indiscriminate radiant energy that can't be 'turned off' or shielded. Thus, the Ford is a constant electronic beacon for enemies! CNO Greenert noted this specifically, and EMCON in general, but did nothing about it.

      The examples are nearly endless.

      Delete
    8. "What I believe they lack is the range-finding ability,"

      This is technically true but practically overcomeable (?!). EO ranging works on triangulation and it doesn't have to be much of an offset. A moving ship can fairly quickly establish range and a UAV, even quicker.

      Delete
    9. " towed/tethered electric drone attached to ship"

      One of the major problems with this concept is that the drone is a locating beacon for the enemy. Of course, depending on the type of drone we might be able to make it stealthy. However, if it radiates (radar) then the enemy can see it from much further away than it can see the enemy and the enemy knows that the host ship is a within a tether length of the drone - an exact location for all practical purposes. It's the flashlight analogy. In real combat, no one is going to want to radiate. Modern combat will be 99% passive. We need to learn how to operate passively as we did in the Cold War.

      So, what you say is technically correct but tactically unwise. This is where an untethered, free-flying drone is useful. It can be sent a hundred miles away from the ship, do its scanning of whatever type, and report back without giving away the ship's location.

      Delete
  7. "Human action based on sensor data is inherently flawed (Vincennes)"

    It wasn't just sensor data that led to the downing of Iranian Flight 655. At the time, Vincennes was in Iranian territorial having chased some Iranian gunboats that fired on one of their helicopters. The Vincennes thought they detected an Iranian F-14 that took off from an airport that operated both civilian and military aircraft. And, before firing on the airliner, the Vincennes was unable to contact the aircraft on multiple military and civilian frequencies. Being in Iranian waters and misidentifying the threat of what they detected contributed to the downing of Iranian Flight 655.

    ReplyDelete
    Replies
    1. "It wasn't just sensor data that led to the downing of Iranian Flight 655."

      The sensor data wasn't flawed at all. Aegis was presenting the correct data. The crew misinterpreted it, seeing what they were conditioned to see. For example, the data unambiguously showed the airliner climbing and yet the crew believed they saw the altitude decreasing (assumed attack profile). The best theory is that they were seeing the decreasing range as the altitude but who knows? The point is, they had clear and correct data but totally misinterpreted it.

      Such examples of humans misinterpreting data are endless. All the data in the world is useless if we can't interpret (or even just read it!) it correctly.

      Being in Iranian waters had nothing to do with the incident. The crew was, apparently unaware of their exact location and, even if they knew, didn't care as they believed (correctly) that they had the right to enter those waters if defending themselves. The presence of the Boghammers certainly added to the overall stress level which probably contributed to the misread of the data but the location, itself, was a non-factor.

      The other ships in the immediate area correctly interpreted the airliner data for what it was but the Vincennes saw what they had conditioned themselves to see - an attack.

      Delete

Comments will be moderated for posts older than 7 days in order to reduce spam.