Using Large Language Models to Protect Satellites from Attack
War has broken out in the Pacific, and our adversaries are using everything in their arsenal to disrupt our satellite communications and surveillance— to strike us blind. They’re trying to jam the signals between our satellites and ground stations. They’re trying to hijack the satellites, by sending commands that seem to be coming from our ground stations, but are actually coming from their own. They’re aiming missiles and lasers at the satellites, and even using their own satellites to take ours out of commission.
Ideally, our satellites would be able to think for themselves, so they could detect and defend against such attacks almost instantly, without waiting for operators at ground stations to analyze the threats and then determine possible courses of action. Having such a “brain” on board each satellite would be particularly valuable in the coming years, when there may be mesh networks of thousands of small DoD satellites— far too many for ground stations to fully monitor.
Defense organizations may soon have the ability to equip their satellites with this level of intelligence. Large language models, a form of generative AI, can develop a contextual understanding of a situation and—mimicking the human brain— make sophisticated inferences and suggest a complex set of actions.
ONBOARD INTELLIGENCE
For example, based on an awareness that war has broken out, or may be about to, a large language model might infer that certain seemingly innocent radio signals actually indicate a probable attack. The model might then execute the defensive measures it has determined have the highest probability of success, taking into consideration not just the adversary’s capabilities, but also how other satellites in the network are currently faring against similar attacks.
And it would do all this without needing to rely on ground stations to detect and analyze the signals, recognize the threat, and then work out how best to respond. It is important to note that any actions suggested by large language models would be constrained by humans through guardrails, based on mission context.
Attacks on satellites—whether by cyber, missile, laser or an enemy satellite—can happen so quickly that instructions from ground stations may not arrive in time. A satellite with a large language model doesn’t have to wait for instructions from a cybersecurity expert on the ground, for example. The large language model on the satellite is the cybersecurity expert.
In a sense, a large language model would be like having a team of human operators on each satellite, performing a number of specialized actions at once—such as analyzing data on an attack, formulating a response, and communicating with other satellites in the network.
And a large language model’s response to an attack can be highly sophisticated. For example, if an adversary fires a ground-based missile at a satellite, the model on the satellite might quickly figure out how to outmaneuver it.
Or, a model might recognize that an enemy satellite is moving into a position that suggests it is about to attack. The model could then deter- mine the best defensive measures— even anticipating how the enemy satellite might respond to those actions, and plotting out moves to outwit it,
like a chess game.
SOPHISTICATED COLLABORATION
With mesh networks, satellites connect with each other through an “internet in space,” and can communicate even if signals from the ground are disrupted. It’s similar to the way Uber works. Each Uber driver serves as a node in a network, providing information to help create a common operating picture. And what one satellite sees, they all see.
If a satellite in the network were attacked, its large language model could not only determine the best defense, it could pass that information along to all of the other satellites. For example, say an adversary jams the ground signals going to a group of satellites. Large language models on those satellites might detect the attack and quickly switch communications to different frequencies, with each model choosing the frequency it predicts will work best.
If a satellite finds a successful frequency, it can communicate that to the others in the immediate group under attack—as well as to the thou- sands of other satellites in the network. If one of the other satellites picks a bad frequency, and is cut off from the ground, it can communicate that to the group as well. The large language models in a mesh network combine what they’ve learned to figure out what works and what doesn’t, as teams of human operators would. With each attack, the network of large language models get smarter about defense.
Just as important, the large language models in the mesh network would work together for the greater good— that is, taking defensive actions not just to protect themselves, but to make sure the satellite constellation as a whole is doing what it needs to do. This might even mean that some satellites would sacrifice themselves— moving into the path of incoming missiles, for example—to protect the larger network.
AWARENESS OF CONTEXT
One of the strengths of large language models, compared to conventional AI, is that they have a much greater ability to understand context. Say, for example, a model learns that the network is under attack from an adversary, and then gets commands from the ground that don’t reflect the conflict, such as an order to observe a region far from the war zone. The model might then take steps to determine whether it is being hacked—it could, for example, query other satellites about whether they are getting the same commands. It might also alert operators at the ground station of the possibility of an insider threat.
Having hundreds or thousands of large language models in a network would help make sure any single model stays accurate and on task. If a model went rogue, so to speak, or was compromised by an adversary, the other models in the network would likely recognize that it was deviating from the group—and possibly quarantine it. They might designate another satellite to take over its role, and perhaps recommend that ground control shut it down.
It would be no more difficult or expensive to equip satellites with large language models than with conventional forms of AI. Large language models are offering new opportunities for defense organizations in a variety of applications—including in protecting satellite communications and surveil- lance from crippling attacks.
LT. GEN TREY OBERING ([email protected]) is a Senior Executive Advisor at Booz Allen, specializing in space and missile defense. He is the former Director of the Missile Defense Agency.
COLLIN PARAN ([email protected]) is an AI architect at Booz Allen who builds large language models for a variety of applications for the Space Force, Navy, Army and Air Force.
Booz Allen subject-matter experts Evan Montgomery-Recht, Timothy Snipes and Karis Courey contributed to this article.
Check out more sponsored articles.
Smart Shipyards Are Essential to Navy SIOP Transformation: A Digital Backbone Is Critical to Force Generation and On-Time Deployment of Ships
The national security imperative for modernization of U.S. Naval shipyards has been well recognized for years. The Navy’s Shipyard Infrastructure Optimization Program (SIOP) Office was created in 2018 recognizing that America’s shipyards were designed in the 1800s and, while there has been some modernization since then, the Navy’s four organic shipyards – Norfolk, Portsmouth, Puget Sound, and Pearl Harbor – require critical upgrades given their vital roles in maintaining and updating nuclear-powered aircraft carriers, submarines, and other ships to ensure their warfighting readiness for the future
The Navy currently has committed $21 billion to SIOP, and while the planned physical infrastructure upgrades are essential (for example, the latest ballistic submarines cannot fit into existing drydocks), many national security experts are concerned that not enough effort and resources are being dedicated to an equally crucial component of a successful modernization program: digital transformation for the 21st century.
One year ago, retired Admiral James Foggo – former commander of US Naval Forces Europe and Africa and commander NATO Allied Joint Force Command Naples – correctly outlined the criticality that any shipyard optimization include a “digital backbone.” Admiral Foggo asserted the reality that the United States will not build as many ships as China for the foreseeable future, and emphasized the need for a more modern and more efficient digital infrastructure to rebuild and maintain our warfighting readiness advantage. The three main challenges to shipyard modernization, according to Foggo, are inefficient workflows, talent & training gaps, and oceans of data. All three can be overcome through a modern digital approach that incorporates additional investment and the establishment of trust from operators on the available new technologies that have been in use in the commercial industry for years.
Foggo’s proposal, accordingly, incorporated four main tenets: a dedication of an additional three percent to the overall SIOP budget to digital transformation objectives, a ground-up approach that involves the engineers, project managers and artisans from the outset and throughout the projects, demonstrating value using existing datasets, and of course, do no harm to the complex environments of a naval shipyard.
Affirmation for Digital Transformation from Leading Shipbuilding Scientists
Admiral Foggo is not alone. In November 2023, Dr. Jong Gye Shin spoke at the American Society of Naval Engineers’ Technology, Systems & Ships / Combat Systems Symposium about the successes South Korea and other nations have achieved in modernizing commercial shipyards. Jong is a three-time winner of the Elmer L. Hann Award (often called the Nobel Prize of Shipbuilding) and was named in 2023 as the chairman of South Korea’s Committee for Expertise of Shipbuilding Specifics. At the symposium, Jong noted both the complexity and differences between various shipyards to make the case for “smart” shipyards, incorporating AI and other modern technologies to manage the volume and complexity of data and eventually even possibly leading to fully automated shipyards.
Jong punctuated the success of this approach by noting that Korean commercial shipyards produce over 200 ships (50,000 tons) per year, while American shipyards produce as few as 10 (10,000 tons). And while those are commercial facilities, Jong also noted the vastly higher number of shipyards and of production capacity the Chinese Navy possess versus the United States. Transformation and innovation are enabled through three lines of effort: top-down, bottom-up, and scalability.
A Three-Pronged Approach To Modernize U.S. Navy Shipyards
The challenges facing the U.S. Navy to realize the visions of Admiral Foggo and Dr. Jong are significant, but not insurmountable. If the resources Admiral Foggo outlined – a small percentage addition to the overall SIOP budget – are allocated, the Navy’s SIOP program can achieve quick, lasting, and scalable results enabled by modern digital equipment.
Achieving this goal begins with an approach along three lines of effort. First, a “top-down” approach is essential for establishing and executing the vision, strategy, and governance for this modernization and includes the necessary program management, organizational change, and governance structures. Equally vital is a fact-based assessment of digital opportunities and a framework for an overall strategic blueprint. Furthermore, this top-down approach demonstrates leadership buy-in to all stakeholders and personnel, which is consistent with Admiral Foggo’s emphasis on the crucial establishment of maintenance and trust. This approach, used regularly in large commercial enterprises, achieves real cultural change, an absolute imperative to achieving real transformation in our public shipyards.
This trust is enhanced by the second line of effort: a “bottom- up” orchestration of pilot programs, prototypes, and learning opportunities running in parallel and simultaneously with the top-down strategy. The bottom-up aspect of the transformation requires establishment, adoption, and execution of Agile and DevSecOps methodologies, proven and trusted software and technology development organizational structures. Agile and DevSecOps place the emphasis on collaboration and constant communication through every step of development, from ideation to creation to execution. The approach also enables the development of minimum viable products (MVPs), which can be released and evaluated incrementally, as opposed to older waterfall methodologies which do not allow for the opportunity to implement, let alone evaluate, segments of work until final release.
This leads to the third line of effort, scalability. This is a projected approach, based on the evaluation of previous releases and MVPs. But Scaled Agile methodologies, by design, incorporate scalability into their project roadmaps. Employing the first two prongs followed by the third not only creates a technological transformation that can be constantly evaluated, but also enables scaling in the quickest manner possible.
Shipyard Digital Transformation: The Essential Path For The Navy
In view of the current capacity of the naval shipyards, the continued and expanding readiness challenges facing the Navy, and the significant investments being made through the SIOP program, it is critical that the Navy include and prioritize the digital transformation of our organic shipyards commensurate with the needed physical infrastructure upgrades. The combined top-down, bottom-up, and scalability approaches to this digital transformation will help the Navy to develop the smart shipyards required to achieve the objectives of SIOP. It is clear that the technology-based optimization of the United States Navy’s shipyards is an achievable objective, for the near and long-term future.
Check out more sponsored articles.
It's a Complex and Dangerous World Out There. DOCA Has Been Providing Clarity Since 1952
The U.S. Civil-Military Divide
Ironically, as our military, intelligence, and foreign services have been engaged in the longest period of sustained conflict in the nation’s history, less than one percent of American adults have served their country on active duty. As our military shrinks, the connections between military personnel and the civilian population naturally grow more strained. This creates a destabilizing divide in our society’s ability to project power precisely when unity of effort is essential for global peace and prosperity. There is increased potential for the United States to experience, in the words of former Secretary of Defense Robert Gates, “a void of relationships and understanding of the armed forces”.
The idea of a civil-military divide is not new to American society. In some respects, military values of unity, subordination, and sacrifice run contrary to the cherished American values of individualism, acceptance, and free expression. This increasing gap, however, has potentially catastrophic implications for our capability to act decisively anytime and anywhere our interests require it. If the public fails to understand the importance and relevance of overseas missions, they may increase pressure on Congress to reduce funding for these operations. Furthermore, the ability to recruit qualified citizens to serve is severely hampered when civilians do not perceive the values of a strong and engaged military.
Why DOCA?
The Defense Orientation Conference Association (DOCA) was founded in 1952 by civilians who had participated in JCOC, a Secretary of Defense initiated and taxpayer funded program. JCOC was designed to address the then new, but growing, civil-military divide by creating a conduit to inform the private sector of the missions and operations of the Department of Defense and the challenges it faces in carrying out its goals. JCOC participants gained first-hand knowledge of the military organizations and defense strategies of our country and were given a frank appraisal of the tasks and problems faced. Early JCOC alumni decided to organize a public effort to continue the important work the SECDEF had begun and thus DOCA was born.
Over the past 70 years, DOCA has evolved from a mostly inside the beltway alumni club into a national, non-political, non-partisan, 501(c)(3) non-profit member association of civilian community leaders with the following objectives:
- To enrich member understanding in matters of national security and international relations under the jurisdiction and supervision of the Departments of Defense, State, Homeland Security, and the Intelligence Community
- To enable members to increase general awareness and understanding in our society by conveying information learned to others in their business and social communities
DOCA organizes four to six field orientation conferences annually. They are held in DoD facilities across America and abroad under DOS auspices where the United States has formal bases or established defense missions. DOCA members thereby gain an integrated view of the military establishment and national foreign policy, the problems
confronting the United States in pursuing its policies, and the economic, political, and military means required to carry them out. DOCA has members from all walks of life and would love to add your voice to ours, the only requirement to join is that approved members must be U.S. citizens with a desire to learn and share.
DOCA Gives Back
- While DOCA has the primary mission to explore, to learn, and to share, we also strongly believe in giving back.
- DOCA provides active support to the National Defense University International Fellows and Counter Terrorism Fellows programs.
Through the DOCA Defense Fund, our members make significant donations to the MWR organizations directly supporting those who serve and their families on every conference trip organized.
FIND OUT MORE AT DOCA.ORG.
Check out more sponsored articles.
How Satellite Swarms Can Take Down Hypersonics
An adversary launches hypersonic missiles at a carrier group in the Pacific. Booster rockets carry the warheads—hypersonic glide vehicles— to the edge of space. But before the boosters can release their payloads, they are suddenly attacked by a swarm of small satellites. The satellites smash into the boosters, destroying them.
The technology needed for such a defense against hypersonic missiles is now available, and it is cost-effective. By leveraging dramatic cost reductions in space launch, as well as a range of technologies currently used in commercial endeavors—from Starlink to Uber—the joint forces have new opportunities to reduce the hypersonic threat. And satellite swarms can also be a powerful defense against conventional ballistic missiles.
Hitting Hypersonics At The Edge of Space
The Defense Department is improving its ability to intercept hypersonics in the glide phase, using both conventional and hypersonic missiles. But that remains an extremely complex task, given the glide vehicles’ speed—more than a mile a second—as well as their formidable maneuverability and low flight paths.
The idea behind swarms of small satellites is to destroy hypersonic missiles when they’re most vulnerable— when they’re in the boost phase, on a predictable trajectory like a conventional missile, and are easier to detect and track. Although it’s difficult for land-and-sea-based defensive weapons to get to the booster rockets before they release their warheads, swarms of small satellites are in a much better position to intercept them. The satellites can track the hypersonic missiles from the time they’re launched, and can be there when the boosters reach the edge of space. The same approach can be used to intercept an adversary’s conventional ballistic missiles throughout most of their trajectories.
How The Swarms Work
A swarm might have thousands of small satellites in low-earth orbit, at the edge of space, something that SpaceX’s Starlink constellation has shown to be both technologically and economically feasible. There are now about 5,000 small Starlink satellites in low-earth orbit, and SpaceX has received government approval to increase that number to 12,000.
The satellites in the DoD swarm would connect with each other through an “internet in space,” a network backbone created by another set of satellites. With this approach, all of the satellites in the swarm have the same information, so that if one satellite spots a missile launch, they all see it. It’s similar to the way Uber works. Each Uber driver serves as a node in a network, providing information to help create a common operating picture.
Another advantage of the swarm’s network is that it is largely protected against jamming. While an adversary might be able to jam a handful of the small satellites, that would have no effect on the overall network. This is another way that the swarm is similar to Uber’s peer-to-peer network: if an Uber driver accepts a rider and then drops out, the system automatically looks for another driver. In this sense, the swarm’s network is
self-healing.
Advanced Autonomy
With the DoD swarm, the network backbone would allow the small satellites to coordinate their efforts autonomously, with the help of AI. For example, if the swarm spots 20 booster rockets, the satellites would decide among themselves which satellites will try to intercept which boosters. They might take a shotgun approach, with five or ten satellites, for example, trying to hit each booster. Once one of the satellites succeeds, the others rejoin the swarm, ready to go after other boosters.
The satellites can also aid land- and sea-based defensive systems. While much of the swarm might be intercepting boosters, other satellites might use lasers to illuminate glide vehicles that make it through.
The artificial intelligence that can provide the small satellites with these and other advanced autonomous capabilities is currently available. And in fact, those capabilities are akin to the ones that will help unmanned surface, undersea and aerial vehicles achieve their own autonomous missions.
Cost-Effective Deterrence
One challenge facing swarms is that, in the booster phase, hypersonic missiles look the same as conventional ballistic missiles. So, the satellites would have to go after every missile— they can’t wait to see which ones are carrying hypersonic glide vehicles. This means that if an adversary launches hundreds of missiles—both conventional and hypersonic—the swarm may be unlikely to get them all. However, the adversary would not know which of its missiles will get through and which ones will be destroyed, potentially leaving open a huge response capability. In this way, satellite swarms can act as a deterrent.
As SpaceX has demonstrated with its Starlink constellation, putting thousands of small satellites into orbit is cost-effective. One reason is that multiple small satellites can be part of a launch, which accounts for a major portion of the cost of a satellite. The swarm’s small satellites are also less costly to build, operate, and maintain.
Hypersonic missiles pose a new kind of threat, requiring new kinds of defenses. One such defense is a satellite swarm that targets the hypersonics where they are most vulnerable—at the edge of space.
Lt. Gen. Trey Obering ([email protected]) is a Senior Executive Advisor at Booz Allen, specializing in space and missile defense. He is the former Director of the Missile Defense Agency.
Vito Partipilo ([email protected]) is a senior lead technologist at Booz Allen specializing in space missile defense architecture modeling and simulation. He is a chief engineer supporting the Space Sensing Command at Los Angeles Air Force Base.
Check out more sponsored articles.
Overcoming Shattered Supply Chains With AI
If supply chains are disrupted during a conflict in the Pacific, commanders will have to figure out—on the fly and often separate from one another— how to get logistical support through other means. However, they may not have the information they need to get that support in the fastest and most secure ways possible. And, if they tap supplies from alternative sources, they may not know how that will impact the missions originally designated for those supplies.
Deep reinforcement learning—an emerging form of AI—may soon make it possible for the joint forces to create what might be thought of as a self- healing supply web. For missions in the Pacific, for example, this web would provide commanders with a constantly updated, near-real-time operational view of supply chains, platforms, and recommended routes.
If, for example, a port in the Pacific were lost, commanders would see a revised operational view showing new locations where the replenishment could come from, and the best ways to get it to the various points of need. The supply web would also map out the cascading implications of the lost port, showing how OPLANs and
missions across the Pacific would likely be affected. Dashboards would allow commanders to work together to choose among the best alternatives, based on mission priorities, rapidly changing conditions, and other factors.
Transforming From Supply Chain to Supply Web
The supply web would be created by securely bringing together a wide range of siloed supply-chain data from across the DoD, and consolidating it in a data mesh. Advanced analytics then use that data to map out the entire supply chains for various OPLANs, operations and exercises. The data includes all the relevant ports, airfields and other supply sources, as well as the current and expected stores of fuel, munitions and other supplies at those locations. Various other aspects of the supply chains are also integrated into the supply web, such as capacity, routes, and expected demand.
The next step in creating the supply web is to bring in deep reinforcement learning, a goal-based form of AI. Deep reinforcement learning uses numerous possible scenarios—created through modeling and simulation— to learn what works and what doesn’t to achieve a particular goal.
A key feature of deep reinforcement learning is that it essentially adopts the point of view of a particular entity— whether a person, a country, or for example, the joint forces and allies and partners in the Pacific—and then tries to achieve that entity’s goals. With the supply web, the reinforcement learning might look at hundreds of thousands of scenarios of how a conflict in the Pacific could unfold, learning not just what will help a particular mission, but what is needed by the allied forces as a whole.
Even before logistics become contested, the supply web would serve several important functions. The supply web’s AI looks at the current supply chain operational view, and evaluates numerous scenarios—far more than human planners could—to identify for commanders optimal ways to support OPLANs and missions.
When there are changes to supply chains, such as shortages or delays in moving shipments, that information is automatically fed into the data mesh. The supply web is constantly pulsing the mesh, staying fully updated.
At the same time, the AI uses intelligence and other information to predict where supply chains are most vulner- able to disruption by adversaries, including through kinetic warfare and cyberattack. Such vulnerabilities might be hidden, and revealed only by working through many thousands of possible scenarios. This gives commanders the chance to address those vulnerabilities before a conflict— for example, by moving supplies afloat or to locations that adversaries may
be less likely to attack.
A Self-Healing Web
The supply web becomes particularly valuable if supply chains are disrupted. If a replenishment ship is lost, for example, the deep reinforcement learning reconfigures—in near-real time—how supplies can get to the points of need. In this sense, the supply web is self-healing. One reason this healing happens so quickly is that the AI doesn’t start from scratch—it uses what it has already learned through the hundreds of thousands of course- of-action scenarios.
In a conflict, the supply web would continuously reconfigure logistics in any number of ways. For example, a multi-day battle in the Pacific might take a carrier strike group
so far afield that many of the ships would run out of missiles before the expected replenishment could reach them in time. The AI would run new scenarios—leveraging its prior learning—to work out the how supply chains could be rearranged so that the group could get the replenishment.
Currently, it is difficult for planners to get a full understanding of how multiple supply chain disruptions in a conflict might cascade across an AOR. Supplies will have to be rerouted in numerous ways, all at the same time. If a carrier gets fuel from a new location, what does that mean for the missions that were originally meant to get that fuel?
The supply web’s deep reinforcement learning would work out these implications in near-real time as the disruptions occur. And instead of reconfiguring supply chains mission by mission, it would work toward the ultimate goal—winning the war.
The supply web would not take away decision-making from commanders. Rather, it would give them more hard data to work with, and help them make faster decisions. For example, the supply web might present one alternative that will get supplies to the point of need faster, and another that will be slower but more secure. Commanders still need to use their experience and judgment to decide on the best path.
Commanders across an AOR can work together on the alternatives through common, interactive dashboards. They are able to ask questions of the data to gain more insight for decisions. And, they can add in new information for the supply web to
consider, such as changing conditions, mission priorities and OPLANs.
Col. Boyd Miller ([email protected]) is a contested logistics subject- matter expert in Booz Allen’s artificial intelligence division. He has more than 30 years of experience in defense, joint, and maritime logistics operations, including as J4 Director, United States Southern Command.
Ki Lee ([email protected]) is Booz Allen’s Global Defense Technology Officer. He drives technology innovation, application, and adoption for Booz Allen’s global defense business, with a focus on supporting mission needs and gaps.
Scott McCain ([email protected]) is a veteran logistics and sustainment expert at Booz Allen who designs, develops, and implements cognitive-enhancing software solutions that enable a wide range of DoD clients to address the complex challenges of contested logistics.
Check out more sponsored articles.
Fighting The U.S. Navy's Slow Creeping Enemy: Corrosion
Corrosion is a billion-dollar problem for the U.S. Navy as constant operations in abrasive environments put its equipment under significant duress. A 2014 U.S. Department of Defense (DoD) report estimated that roughly $3 billion is spent annually fighting corrosion on Navy ships and vessels—nearly one-quarter of its overall maintenance expenses. And these costs aren’t limited to maritime equipment—between 2017 and 2020, the U.S. Navy spent more than $2 billion on corrosion maintenance for its F/A-18C-G fleet alone.
Independent nonprofit science and technology research organization Battelle has helped the DoD identify and address many corrosion challenges. Its multidisciplinary, comprehensive materials science and engineering expertise is a one-stop-shop to solve the most challenging material problems. In its well-equipped facilities, Battelle performs systematic material characterization assessments, material development, performance test and evaluations, and specific failure analysis. When Battelle’s full capabilities are brought to the table, myriad challenges can be solved to minimize maintenance costs and ensure critical systems are always mission ready.
Fighting Corrosion Requires Tailored Solutions
The U.S. Navy’s equipment and assets execute different missions in diverse environments, so there is no one-size-fits-all solution to its corrosion challenges. Operating environments for Navy surface ships, submersibles, and aircrafts vary significantly, depending on location, usage, the materials involved, the age of the equipment and more.
Battelle has multiple indoor and outdoor facilities for research and development, characterization, and the testing and evaluation of advanced materials. The laboratories at Battelle’s Center for the Characterization of Advanced Materials (CCAM) are explicitly designed for coatings research and development and have advanced characterization systems for spectroscopy, microscopy, failure prediction, and analysis. The Center also houses equipment for accelerated weathering and cyclic exposures, which enables Battelle to evaluate the impact of various environmental conditions on specific materials.
At its one-of-a-kind Florida Materials Research Facility (FMRF), Battelle can conduct outdoor materials characterization to assess atmospheric and marine exposure of materials. (see Figure 1.) It’s located in one of the most corrosive environments on Earth and the property includes the only commercial oceanfront facility for subtropical exposure studies in the U.S. It’s also equipped with an array of exposure racks and fences that conform to American Society for Testing and Materials (ASTM) and two acres of beachfront exposure area less than 70 meters from the Atlantic Ocean. This means Battelle can expose sample materials to nature’s harshest elements. Battelle’s FMRF also boasts access to tropical and subtropical marine fouling organisms with a level of biodiversity that is not achieved further north or south. Battelle can test marine materials, antifouling coatings, and foul-release products in fully submerged, floating waterline and splash zone exposures on its marine immersion dock in the harshest of fouling conditions available. With the ability to perform both ASTM and Military Spec testing, Battelle can support failure analysis, material evaluation, and qualification against the most rigorous standards.
Battelle’s extensive failure analysis expertise also enables it to identify key material properties and conduct analysis of alternatives (AoAs) to achieve system requirements. For example, it has assessed coating vulnerability for specific applications and environments and proposed formulation enhancements to ensure the coating’s integrity and ability to protect the substrate.
Additionally, Battelle has a team of subject matter experts evaluating impacts of fungi, bacteria, and other biological growths on coatings used on a variety of substrates within the DoD. As part of a recent U.S. Air Force project, Battelle worked with a manufacturer to improve the weathering performance and corrosion resistance of a powder coating used on aircraft ground support equipment operating in hot and humid conditions.
Combating the Slow Creep of Corrosion
Battelle is equipped to help the U.S. Navy stay ahead of the corrosion and mitigate corrosion risk to its assets and equipment during deployment and make informed acquisitions and sustainment decisions. For example, Battelle successfully evaluated Supersonic Particle Deposition (SPD), or Cold Spray, a novel additive technology to mitigate galvanic corrosion of NAVAIR gearboxes.
Figure 2: Battelle’s marine dock provides exposure to the most aggressive of marine environments.
Proactive planning and prevention mitigates corrosion risks and can save millions of dollars. For the U.S. Navy, Battelle’s unmatched expertise and ability to offer life-cycle management solutions across multiple systems is particularly germane. These solutions extend the service life of vital assets, improve mission readiness and reduce the total ownership costs—a true win-win scenario.
Check out more sponsored articles.
Why We Need "Brain-Inspired AI" For True Unmanned Autonomy
A group of five unmanned surface vehicles in the South China Sea spots a contingent of enemy vessels, but can’t get that information back to operators— it’s a contested environment, and satellite communications in the area are jammed. The UVs, working together, determine that one of them needs to leave the area to send a message back.
They decide among themselves which of the five should go, based on which has the best information and the best chance of sending the message without being detected. The chosen UV leaves the area, and figures out for itself when conditions are right to send the message, and the safest, most efficient way of sending it.
The artificial intelligence that can provide UVs with these and other advanced autonomous capabilities
will soon be available. But there’s a problem. Such sophisticated AI requires computers that are too big, and require too much power, to fit on UVs.
What the AI needs is a way to lighten its workload, so that onboard computers can be smaller and use less power. And two new approaches are now able to do that, by making computers—and the AI itself—mimic how the brain operates.
One approach is an emerging new design for computers, allowing them to process and store information the same location—similar to the way the brain does—rather than in two different locations. With the second new approach, the AI reaches conclusions with less data, through inference—comparable to how we can identify an object even if we have only a partial view of it, by filling in the blanks.
Currently, small, low-power computers on UVs can only support “narrow AI”—good for a few basic activities, such as surveillance and reconnaissance. But with the two “brain-inspired” approaches, even highly sophisticated AI can run on the smaller computers. This makes it technologically feasible for the joint forces to bring high-level autonomy to unmanned surface, undersea, and aerial vehicles in the Indo-Pacific
and beyond.
Beyond Narrow AI
With narrow AI, unmanned vehicles are not intelligent enough to act autonomously in a number of important ways. For example, they can’t independently determine whether something they’ve spotted is important enough to alert an operator—currently, UVs check in at scheduled times. They don’t always know how to use their fuel efficiently when tracking contacts, or how to conduct ISR without being detected. They typically can’t autonomously distinguish between combatants and non-combatants, and don’t know how to apply rules of engagement. They have only limited situational awareness.
UVs theoretically could tap into sophisticated AI by connecting to the cloud—but that’s not a workable option. UVs can’t count on satellite communications in a contested environment. And power and bandwidth constraints would limit back-and-forth with the cloud, even in peacetime. So, the AI has to be able to run onboard.
Mimicking The Brain
The two brain-inspired approaches don’t make the AI smarter—AI is already gaining the ability to provide many aspects of advanced autonomy. What the approaches do is simply make it possible for the AI to run on the small, low-power edge computers that UVs have to rely on.
One of the approaches actually changes how computers work. Today’s computers have separate processing cores and memory cores. This means that for each computation, the processor reaches into memory, takes out the data it needs, and then brings it back to process it. That continuous back-and-forth makes for a heavy workload—particularly for AI that does billions of computations a second. While the back-and-forth may not be a problem on large, powerful computers—such as those on conventional ships—it can quickly overwhelm a UV’s edge computer.
Our brains operate in a different way. We’re able to hold much of our memory in the same place that we process information, which allows even our most complex thinking to be almost instantaneous—a definite evolutionary advantage. Mimicking the brain’s design, AI researchers are developing computers that put processing and memory in the same place. This makes the workload of even sophisticated AI manageable on a UV computer.
Training AI To Use Interference
Another way to reduce the workload is to simply use less data. AI researchers are achieving this by mimicking how the brain uses inference to make sense of the world with limited information. For example, when we’re driving, we can anticipate the actions of other drivers by subtle cues, such as a car speeding up before changing lanes, or a car edging to the left as it approaches an intersection, in advance of actually making a right turn. We’ve seen these scenarios so many times that we don’t need any additional information to adjust our driving. Our ability to make inferences and predictions from just a few cues is one reason why we can (usually) drive safely on auto-pilot, our thoughts elsewhere.
By training AI to infer from a few cues, researchers are greatly reducing the amount of data—and power—the AI needs. For example, the AI might be provided with the “pattern of life” of an adversary’s vessels in a particular area. If the UV’s sensors pick up an anomaly—such as a vessel that’s in an unexpected location, or is behaving in an unusual way—those may be cues that will enable the AI to infer the vessel’s intention. The AI doesn’t have to piece together every detail about the vessel, or sort through every potential action it might take. By picking out only the relevant cues, the AI could reach its conclusions with just a small fraction of possible computations—making it workable on a small edge UV computer. And the AI would be just as accurate as AI running a large, powerful computer on a destroyer.
Training AI to use inference is both an art and a science. The ability to select the right cues, and fully understand their implications, requires extremely deep domain and mission knowledge. At the same time, AI experts need know how to apply that knowledge to achieve autonomy.
If unmanned vehicles in the Indo-Pacific and elsewhere are to gain the level of autonomy required by the joint forces, AI-enabled edge computing needs to be rethought. The human brain can provide the inspiration.
Ayodeji Coker ([email protected]) is a former senior leader at the Office of Naval Research who is now an executive advisor at Booz Allen, where he leads intelligent autonomous systems strategic initiatives for the Navy. His roles at ONR included portfolio manager for autonomy, and leader of the Navy’s intelligent autonomous systems strategy.
Jandria Alexander ([email protected]) is a vice president at Booz Allen who leads the firm’s business for NAVSEA and S&T, including unmanned systems, resilient platform and weapon systems, data science, and enterprise digital transformation strategy and solutions for Navy clients.