Unmanned and autonomous systems will play a significant role in tomorrow’s conflicts. Great leaps have been made in autonomy, and huge potential still exists for development in the underwater domain. The United States and United Kingdom have the technology today (or available in short order) to develop a mostly autonomous platform capable of filling many of the missions held by nuclear attack submarines (SSNs).

Scene Setter

The year is 2030 and tensions are heightened between NATO and an aspiring great power nation with a large maritime border, herein called ‘Country Orange.’ The springtime sun’s daily heat eventually dissipates the binding sea ice. A ballistic missile submarine (SSBN) gets underway from Country Orange’s fleet base. From deep inside the SSBN, several pieces of machinery create specific sounds that radiate into the cold water. Thirteen nautical miles off the coast of Country Orange, a collection of transducers, processors, and propulsion equipment inside a smooth metal casing lie in wait. The SSBN’s sounds reach the device’s transducers, which relay the detected energy levels to the onboard processors. The processors analyze the specific combination of sounds and, like a key in a lock, the sounds betray their source to the enemy. Having spotted this target of interest, the autonomous undersea reconnaissance vehicle (AURV) gathers data and, after a period of analysis by processing algorithms, confirms the SSBN’s identity. Covertly, the AURV distances itself from the SSBN before coming shallow, extending a small radome above the ocean surface, and sending a brief encrypted signal with the collected data on the SSBN in a narrow beam to a distant satellite. The AURV then slips into the frigid depths and gets in trail.

General Construct

The Mark 48 ADCAP torpedo provides an excellent template for the general scale of technology required for an AURV. ADCAPs are, in essence, small autonomous submarines. The heavyweight torpedo has many of the components necessary to conduct basic submarine missions: sonar, fire-control processing, and propulsion.[1] Given the expected service life and broad mission set of an AURV, the sensor and propulsion components would need to be more capable than those of a torpedo. Furthermore, the propulsion plant required for the endurance of an SSN would be larger than that of an ADCAP. Keeping all this in mind, an optimally-sized AURV would have a relative size between an ADCAP (19 feet x 1.7 feet) and Boeing’s extra large unmanned underwater vehicle (XLUUV), the Echo Voyager (51 feet x 8.5 feet).[2]

Sonar

The combined capability of an AURV sonar suite ought to be analogous to that of an Astute– or Virginia-class submarine.[3] It would not require a great technological advance for the bow array of a U.S. or U.K. platform to be scaled down and installed in the nosecone of an AURV. Similarly, the inclusion of side-mounted sonar (such as a wide aperture or flank array), or a high-frequency active array would be an easy and beneficial addition to the design of a future AURV. Somewhat more difficult would be designing a UUV-sized towed sonar array. There are multiple designs for such arrays, however the tow cables and transducers constitute a significant volume that may be prohibitive for inclusion on an AURV. To mitigate the volume problem, a ‘clip on’ array, rather than a ‘reelable’ one may be used. Alternatively, a similarly capable hull-mounted array may be developed eliminating the need for a long length towed sensor.

Electronic Support and Optical Sensors

As much as submarines rely on their sonar sensors to execute antisubmarine warfare (ASW), they also rely on their passive electronic support (ES) and optical sensors to conduct surveillance and reconnaissance. The Virginia and Astute programs demonstrated a great leap in automation from their predecessors—using optronics (also known as photonics) masts instead of traditional submarine periscopes.[4] Given that optical sensor data can be collected via a mast and transmitted directly to the combat control system, such a sensor could easily be installed on an AURV. Perhaps more useful given the platform’s size would be a buoyant sensor that could be deployed from a reel onboard the unmanned unit. This would make for easier depth-keeping and allow for a smaller above surface profile. On this same buoyant component, an ES sensor could be attached, and its data could be transmitted and processed similarly.

Combat Control

The “brains” of the entire AURV platform will be its combat system. The development of technical insertions in both the Astute and Virginia platforms has allowed rapid modernization of combat systems. More advanced processing equipment has supplied warfighters aboard these platforms with additional tools for interpreting the data coming in from sonar, ES sensors, and optronics. Much of the process of analyzing sonar data and creating a solution (where the target is and which way its moving) has already been automated. Within the Royal Navy and U.S. Navy, capabilities for automated processing of optical and ES data are behind those of sonar, but with a significant enough investment of effort in the space of two to three years, the United States and United Kingdom could have a combat system that ingests all of these data types, and develops target solutions without the assistance of a human operator. Much of the development and assurance of the system could be handled through robust computer simulations. Thousands of trials could be generated with randomized target contact data. After each “run,” the generated system solution could be compared with the target simulation and the artificial intelligence algorithms adjusted as necessary. This process would provide for development of a strong and intelligent automated combat system.

To ensure that deployment priorities are executed in the austere communication environment of an SSN deployment, units could be programmed with prioritized tasks. There should be a process for AURVs where they would be similarly programmed or reprogrammed throughout the course of a deployment to account for their tasking guidelines. This could occur periodically while the unit is in a passive or active communications posture. With this programming in place, the AURV can decide how it will be employed based on what is held on its sensors.

Propulsion

The most pressing challenge to the AURV concept would be the design of its power plant. Even today’s advanced UUV batteries would not provide sufficient power for a unit to conduct a multiweek deployment without supplemental charging. While it would certainly be possible to develop an at sea charging capability, either via an auxiliary ship or on the seafloor, this would not be practical given the area denied environments an AURV would be tasked to operate in. The Caterpillar diesel engines installed on board Virginia-class submarines are smaller than the models on board Los Angeles-class submarines and are, in part, operated via a touch screen. It would not require a huge investment to create an autonomously run scaled down diesel engine for use in a future AURV. In fact, autonomous diesel technology has already been tested on board the Echo Voyager . Couple this concept of a smaller diesel engine with advances in air independent propulsion, and you have a reliable, long duration, and air-independent source of power.

The fundamental technology for an autonomous nuclear reactor exists. Control rods in a reactor, ie: devices which moderate the fission process, already are electrically controlled, and valves of many types can be operated electrically. Central to the design of autonomous nuclear power plant would be a processor that coordinates movements of control rods and valves via based on changing power plant conditions. The steepest hill to climb in nuclear automation will be building public trust in the design, which is not a trivial obstacle. A great amount of research and development would be required to ensure that an unmanned power plant would have a hyper-conservative fail-safe design. All these points aside, the concept of an autonomous nuclear reactor would be greatly beneficial for an AURV.

Challenges

As with any computer-driven system, hacking and malicious software installation are a formidable threat. The thought that an AURV could simply be ‘hijacked’ before or during a deployment and steered toward Country Orange’s port is startling. While newer warships have many automated features, there is normally a human in the loop who can regain control if the system does not respond as directed. Prior to bringing a future AURV online, intensive information assurance protections must be in place.

What happens if an AURV is captured? In the short history of unmanned aircraft, there are multiple instances of platforms falling into the charge of an undesirable actor. While an AURV cannot defend itself with an M9 handgun, it may act to minimalize national embarrassment. To do this, its on-board programming must be able to recognize when it has been captured and appropriately destroy sensitive components of its internals.

There is no clear path, as yet, to arming AURVs. Humans cannot yet be completely removed from the kill chain. The denied environments in which AURVs would operate necessitate removal of human operators in the loop. Therefore, it would be wise to table the autonomous weapons discussion until the viability of AURVs as a reconnaissance-only platform is proven.

Conclusion

Both the United States and United Kingdom are investing resources into multiple undersea autonomous platforms. What direction must unmanned undersea vehicles proceed in? It is necessary to focus efforts more narrowly—ideally to a single platform capable of multiple missions. To that end, I suggest the following lines of effort:

1. Develop small but capable sonar systems. Scale down kits from current SSNs, install and test them on board existing UUVs. Ideally, the new sonar systems do not represent a degradation in current abilities. For this reason, the specifications of the new equipment should be similar to their larger SSN counterparts.
2. Build smarter artificial intelligence processing for combat control systems. Through advanced computer modeling, the United States and United Kingdom could pit a notional future automated combat control system against a slew of scenarios (incorporating all types of data—sonar, ES, visual, etc.), and adapt the programming as necessary based on performance.
3. Develop AURV-compatible ES, visual, and external communications systems. The artificial intelligence for image recognition, particularly in the maritime environment, is ripe for further development.
4. Choose an appropriate platform to house these components, and put that weapon to sea.

On either side of the Atlantic, there are projects proceeding on different UUV platforms. The blue water navies of the US and UK are propelled by platforms with a long striking distance. If UUVs are to fight tomorrow’s maritime battles, the new platforms must capable of proceeding underway from Devonport or Norfolk and taking the fight to the enemy’s home shore. Therefore, UUV development efforts must drive toward the end goal of a single multimission platform with a high level of automation and a long range.

Endnotes

1. “Mk 48 all Mods – ADCAP/AT/CBASS,” Jane’s Weapons: Naval 2019–2020 (IHS Markit, 2019), 438–41.
2. “Mk 48 all Mods – ADCAP/AT/CBASS,” Jane’s Weapons: Naval 2019–2020
3.  David Ewing, “Astute class,” Jane’s Underwater Warfare Systems 2011–2012, (IHS Jane’s, 2011), 111; Ewing, “Virginia class,” Jane’s Underwater Warfare Systems 2011-2012, (HS Jane’s, 2011), 133.
4.  Ewing, “Astute class” Jane’s Underwater Warfare Systems 2011–2012; Ewing, “Virginia class,” Jane’s Underwater Warfare Systems 2011-2012, (HS Jane’s, 2011).