Autonomous Undersea Vehicles: An Emerging Intelligence Collection Platform
Unmanned Undersea Vehicles (UUV) have been in development since the late 1950s; they have seen limited military application with the most notable application being in the conduct of mine countermeasure (MCM) operations.¹ As the U.S. transitions to a “Great Power Competition” footing, the Navy finds itself facing adversaries that have both “blue” water and “brown” water capabilities, as well as potential Subsea and Seabed Warfare (SSW) capabilities, which pose serious anti-access challenges to the Joint Force. Although maritime denied-area intelligence operations have long been part of the submarine force mission set, advancements in UUV technology puts them in position to better face these challenges to enable “all-domain maneuver warfare.”² Specifically, UUVs will be instrumental in SSW, Anti-Access/Area Denial (A2/AD) Intelligence, Surveillance, and Reconnaissance (ISR), and Initial Preparation of the Environment (IPOE) for undersea operations.
UUVs can be separated into two distinct categories regarding their autonomy—Autonomous Undersea Vehicles (AUV) and Remote Operated Vehicles (ROV)—and four categories regarding size of hull diameter: small (3-10 inches), medium (10-21 inches), large (21-84 inches), and extra-large (>84 inches).³ ROVs require a tethered system to the launch platform, and as such play a reduced role in future intelligence operations. As a result, this paper will focus on AUVs. UUVs of all sizes can and will play a role in intelligence operations; however, mission sets available will be size-dependent based on payload capacity. Payload loadout is directly associated with UUV class, with large and extra-large capable of carrying multiple subsystems, and having capabilities comparable to manned assets.⁴ Potential launch platforms are also dependent on UUV size. Small and medium format UUVs can be launched via manned submarine torpedo tubes and loadable launchers, while large UUVs can be deployed from Dry Deck Shelters or payload tubes on submarines, and extra-large UUVs must be launched by shore or ship.⁵ Outside of the payload/sensor packages installed, UUVs have three main subsystems, all of which are at various levels of development but are vital for future operations. Those subsystems are navigation, communications, and autonomy.
One of the most important subsystems for the successful operation of UUVs is navigation. Because of the attenuating properties of water, UUVs are unable to maintain constant GPS.⁶ For the UUV to determine its location in relation to the earth, the navigation system will first ingest the initial position from either GPS or the onboard navigation system of the launch platform. Once deployed the UUV’s navigation system will use an Inertial Navigation System (INS) for electronic dead reckoning to estimate its position. INS uses a minimum of three accelerometers and three gyroscopes.⁷ The accelerometers inside the INS measure changes to relative motion to predict the change in position compared to a reference frame.⁸
As the INS predicts the vessel’s position, there will be a build up of error, that adds to the uncertainty of absolute position. While this effect can be minimized with some processing and corrections based on set and drift, it will be necessary to periodically re-fix the vessel’s position with an external source. GPS and bathymetric are two main methods that exist for this, but due to the nature of many of the mission sets for UUVs, bathymetric fixes are unlikely to work. GPS fixes will be possible if the size and payload of the vessel contains the required mast and antenna layout; however, the vessel will have to be able to detect potential jamming and spoofing operations that could be detrimental to fixing position. Because of these issues, the length and type of missions may be limited based on a threshold of position uncertainty.
There are two communications methods available for AUVs: radio/satellite and acoustic.⁹ Each of these systems have significant pros and cons, and the choice to go with one over the other will be largely mission based. Additionally, for non-critical intelligence needs, UUVs can also be recovered, and their storage devices removed for analysis. Effective communications are important as it allows for mission planning updates, multi-platform coordination, and offloading collected intelligence for analysis and decision-making.¹⁰
For high-data-rate transmissions UUVs must rely on radio and satellite communications.¹¹ This can be done either via point-to-point communication between the UUV and a surface or air asset, or it can be accomplished using a global communications system via satellite connection. Both of these communications methods require the UUV to be near surface level, increasing the probability of counter-detection. This risk may be warranted depending on the mission; however, this form of communication also lends itself to potential interception and jamming.¹²
For data transfers between UUVs and other submerged assets, an available but low-data-rate communication solution would be through acoustic communications.¹³ Acoustic communications are possible as the transmitting unit uses modulated sound waves as carrier waves to transfer data to the receiving unit. The receiving unit then demodulates the signal to extract the data.¹⁴
To gain the most effective use of UUVs, the human element must be removed as much as possible from the decision loop. While true autonomy may not be possible, as humans will need to develop the boundaries in which the UUV operates, once the platform is deployed it is imperative that it can conduct the mission with little input.¹⁵ To reach this level of autonomy, it is necessary to further develop machine learning and Artificial Intelligence (AI) algorithms.
Machine learning focuses on “large-scale data for pattern matching and inference,” for the system to better understand its operating environment.¹⁶ Machine learning algorithms have thus far relied heavily on supervised learning in which the system is provided large, labeled data sets to teach the algorithm how to detect and classify objects.¹⁷ This has a major pitfall in that real world detection and classification scenarios may not have sufficiently large, labeled data sets to initially train the algorithm. This is especially true for detection of novel adversarial systems. Outside of detection and classification, which will assist the UUV on making decisions based on a programmed decision tree, autonomy also plays a key role in basic functions. These functions include track planning, collision avoidance, and station-keeping.¹⁸
The U.S.’s near-peer competitors have steadily increased their seabed planning and operations over the last several years. Russia’s manned special purpose submarine, Losharik, is suspected of having capabilities to target deep water fiber-optic communications cables, and the People’s Liberation Army (PLA) Navy is planning an “Underwater Great Wall” of seabed sensors to aid in ASW operations.¹⁹ The U.S. Navy has limited capabilities to effectively respond to these challenges; however, UUVs can be used to fill this capability gap. While the sensors that would be employed for this could be used by manned submarines, seabed systems have the potential to be in both extremely deep and extremely shallow water, which are either unreachable or extremely dangerous for manned systems.
Medium to extra-large UUVs equipped with multi-aperture side scan SONARs and underwater camera systems can be used to generate high-resolution maps of adversarial seabed systems. These systems have already been field tested by imaging shipwrecks.²⁰ This sensor package, coupled with advanced machine learning, will be able to identify objects within the system on the sea floor. Additionally, the same technology could be used to monitor subsea communications lines and detect indications of tampering. For large and extra-large UUVs, once a seabed system is detected, it will be possible to deploy smaller systems to either neutralize or further monitor the system.
Coastal A2/AD ISR
Coastal ISR is another mission set that could be accomplished by manned systems, but in which UUVs are better suited in many cases, as coastal ISR often requires operation in shallower water with higher contact density. Additionally, unmanned systems do not have the housekeeping requirements of manned systems that require coming off station. After being launched by either a submarine or surface vessel, the UUV would enter the designated operating area and begin station keeping. Once on station the UUV would use variations of the same masts and collection systems used on manned submarines to collect intelligence.
Specifically, the UUV would be tasked with the collection of SIGINT from shore-based facilities and vessels entering and leaving ports. A key concern in future conflicts with near-peer competitors is their potential use of shore based anti-ship missile systems. A UUV equipped with SIGINT collection capabilities would be able to map the ELINT environment, and with machine learning algorithms, it could detect and classify targeting and other RADARs. After completing the assigned collection task, the UUV could either be retrieved or it could send off its data package via satellite communications so that the intelligence could be analyzed and fed back for future planning and operations. UUVs would also be able to use these SIGINT capabilities to collect communications signals in the coastal region providing further insight into operations in and around shore-based facilities. In addition to SIGINT capabilities, UUVs deployed in coastal ISR missions would be equipped with a photonics system for capturing full-motion video (FMV) and object avoidance. While the FMV captured by the UUV would not be useful for quick turnaround operations such as in Unmanned Aerial Systems (UAS), it would be useful for establishing operating patterns on the shore as well as visual identification of adversarial ships and systems.
Initial Preparation of the Undersea Environment
Although coastal ISR fits within the category of IPOE, it primarily would be in support of future air and surface operations; a different set of sensors would be required to conduct IPOE for sub-surfaced operations and can be applied to both open ocean and littorals. Submarine operations and acoustic performance can vary drastically across water columns due to changes in temperature and salinity. These environmental factors along with bottom topography can affect acoustic sensor performance, requiring alterations in operations to either search for or evade enemy ASW platforms. While submarines can gather environmental data in-situ to use in operations as well as review historical data, UUVs would extend their reach and gather intelligence on denied areas improving tactical decision-making.
After being deployed from the chosen launch platform, the UUV would proceed to the area of operations and commence environmental analysis. The UUV would then station keep in the water column or rendezvous with the manned submerged asset and transfer environmental data via acoustic communications. That data would be ingested into the onboard environmental analysis systems of the submarine to facilitate operational planning. Depending on the range and endurance of the unmanned system, it would be possible to send it forward to denied operating areas to gather intelligence on the environment and enemy activity to significantly reduce the amount of area familiarization conducted by the manned submarine.
The emergence of UUVs has the potential to fill a major intelligence gap in the form of seabed warfare, as well as replace manned submarines in ISR missions while also contributing directly to their tactical decision-making process. To fully reach that level, the technical hurdles of refined navigation systems and improved autonomy must be overcome. Current communication systems are adequate; however, to get intelligence into the hands of decision makers faster, data rate capacity must be increased. While UUVs of all sizes have a role in the future of subsea intelligence collection, the most capable platforms will be large and extra-large UUVs as their potential sensor loadout can rival that of manned systems.