In the same spirit as my previous essays “What would a robot aircraft carrier look like?” and “What would a robot tank look like?” I’d like to examine what a robot submarine would look like. Specifically, an attack submarine, which is a sub designed to destroy surface ships and other subs (though sub-on-sub combat is INCREDIBLY rare). And by “robot sub,” I mean it is fully autonomous, with only robots and machines aboard and no humans.
And since there are many different kinds of attack subs in service across the world, I’ll focus on just one: the Virginia-class. This is the backbone of the American submarine fleet, and they’re among the most versatile and advanced in the world.
Regardless of country of origin, modern attack subs like the Virginias all share some basic design features: The vessels all have a cigar-shaped hull and a stubby wing-shape called the “sail” sticking up vertically. The rearmost 30 – 40% of the hull is dedicated to propulsion and fuel. The frontmost 10 – 20% of the hull is dedicated to the sonar and torpedoes. The remaining middle part of the ship is for the human crew and their needs, and consists of spaces like the control room, galley, kitchen, bunk rooms, bathrooms, and laundry room.
All of those rooms for humans could be deleted, allowing the sub’s overall length to be shortened by at least 20%. A shortened Virginia would be faster, nimbler and longer-ranged as a result thanks to reduced weight and drag: fluid dynamics shows that subs have the least drag when their breadth-to-length ratios are between 1:6 and 1:8. The latest Virginia-class subs are 34 ft wide and 377 ft long, a ratio of 1:11. Deleting the middle 25% of a Virginia would make its ratio 1:8.3, which is close to ideal.
An automated attack sub would still need robot “crewmen” to do maintenance and repairs, and as such would still need open spaces around important ship components like the engine and torpedo launchers that were spacious enough for the robots to move around them and access them. Additionally, the rooms would need to be connected to each other so robots could move around to meet the needs of the moment. As a result, the basic internal layout of a robot sub would be recognizable to a human submariner.
However, since robots could be created in any shape or size and could be programmed for any environment, many aspects of the sub interior would be very different. Because it is built around human ergonomics, manned Virginia-class subs have three decks (plus a curved storage space under the lowest deck), and on all parts of the ship the ceilings are high enough even for tall people. Without humans, things could be very different. The decks could vary in number and height from one compartment of the vessel to the next. Some might be too low for an adult human to walk through them.
This wouldn’t be a problem since robots of different shapes and sizes could be made for different parts of the sub: Robots the sizes of humans and cats could work alongside each other, and even tight, irregularly-sized areas would be accessible to them. Machinery, tools, cargo, and pipes would be arranged more space-efficiently than it is in human-crewed subs, and I imagine the interior being a bit like an ant colony.
Weird deck layouts could also be possible to maximize the use of space. Instead of three, perpendicular decks with the same ceiling-to-floor heights, imagine a much more complex arrangement where different parts of the ship have different numbers of decks. A human would struggle to keep track of this, but machines wouldn’t. Some of the decks might not even be flat.
The more efficient use of internal space could let us reduce sub’s volume without sacrificing anything. The shape of the hull could be transformed from the straight, cylindrical cigar-shape to a more streamlined shape that reduced drag. The rear half of the sub would taper down more sharply until it ended as a point where the propeller was. Internal volume would be lost in the rear, but the sub could still retain the same power and weapons.
As for the robots themselves, I can only be sure that their limbs will all be designed for grasping, like those of orangutans or spiders. In small, confined spaces, human legs are much less useful than a second pair of “bottom arms.” Filling the interior of the sub with metal gratings, handholds and ladders will make this body form even more useful. They could climb the walls and even ceilings to access everything, allowing better use to be made of the sub’s interior space.
For obvious reasons, they would also be waterproof to great depths. A crew of them could continue working even if their sub sank, and they might be able to fix and refloat it.
The sail (often incorrectly referred to as the “conning tower”) could probably be made shorter, or even totally deleted, in an automated sub. The structure primarily serves as a lookout post and a place for submariners to shelter from the elements when their sub is on the surface. If automated, the sub could watch its surroundings on the surface by extending its periscope, and if its robot crewmen had to do something outside, they would be much less affected by rain, waves, and temperature extremes than humans. Attaching themselves to steel cables while working outside would probably be all that was needed to ensure their safety.
A sail greatly increases a sub’s drag and creates a “wake” behind it that interferes with the propeller’s rotation. Therefore, deleting or at least shrinking the sails should boost the sub’s top speed and fuel efficiency.
The lack of a sail would also make it easier for an attack sub to roll over like a log or like a football spinning through the air. This would let it make tighter turns, which would be useful in combat or when evading enemy attacks. Anyone who has turned around a curve too quickly in a car and felt it dangerously tilt is familiar with the effect of inertia in this context.
While a rollover is disastrous on land, it wouldn’t have to be underwater if the vehicle didn’t make contact with any objects while spinning. Of course, even if a manned sub were technically capable of such a maneuver, it would wreak havoc since crewmen would be thrown around inside. Conversely, on an autonomous sub, the central computer could wirelessly alert all of the robots to upcoming rollovers, giving them time to brace themselves against surfaces and to grab handholds.
Measures would also need to be taken to ensure all the machinery and cargo was properly secured so it wouldn’t be thrown around inside the sub during rolls, but this is easy to do. A bigger problem would be dealing with the service interruptions in pieces of machinery not designed to operate inverted or on their sides, like the nuclear reactor and backup diesel generator. However, given the rarity of rolls (a sub would only need to turn very sharply in emergencies) and their short durations (a handful of minutes per sharp turn), it might not be an issue. Adding a backup electric motor and batteries that would kick in during rolls might be all that is necessary. (To be clear, an autonomous sub would still have a preferred “right side up” that it would be designed around, and under normal conditions it would be oriented in that way.)
Not having humans aboard also removes the need to keep the submarine interior full of oxygen. This is important since oxygen is corrosive and flammable. A robot sub would probably be filled with pure nitrogen gas instead because it lacks those qualities. As an inert gas, nitrogen protects computer chips, which would have obvious benefits for the robots and other machinery.
If, by the future date when autonomous attack subs are being built, backup diesel generators are still being used, then they will need oxygen. The generator room might therefore be the only place in the sub with an oxygen atmosphere.
Putting this all together, what does our robot attack sub look like? It would be a Virginia-class sub, but shorter, sailless, and with a visible taper from the middle of the ship to the propeller. The two horizontal steering fins on either side of the sail would be relocated to the sides of the hull. The robot sub would look like a hybrid of the unbuilt Soviet “Project 673” and American Conform-class attack subs, shown below.
The autonomous Virginia would have the same firepower, power plant, sensors, and stealth features as its manned variant, but it would be faster and more maneuverable, giving it an edge in combat and allowing it to attack targets across a larger geographic area. The deletion of the human crew would also greatly increase the ship’s mission endurance beyond the current 90 – 120 days, which is how long it takes for the food to run out. Thanks to the practically unlimited amount of energy provided by the nuclear reactor, an autonomous Virginia-class sub would only have to return to port if it expended all its weapons or had a serious mechanical problem.
The unmanned subs wouldn’t need to devote time to training missions since their central computers and robot crews could be reprogrammed in minutes as needed. The result of all of this would be a sharp improvement in submarine readiness rates and efficiency. Unmanned subs could patrol the same geographic area and do the same number of missions as a larger fleet of manned subs of the same type.
Links
- To minimize drag, a submarine’s width-to-length ratio should be between 1:6 and 1:8.
https://www.marineinsight.com/naval-architecture/introduction-to-submarine-design/\ - It’s bad news if a submarine rolls over, but it’s probably a solvable engineering problem.
https://www.quora.com/What-will-happen-if-a-submarine-rolls-upside-down-Will-that-ever-happen - A history of sailless submarines.
http://www.hisutton.com/Sailless_Submarines.html