Darpa Kill Web
It is 20XX. A limited war breaks out involving a territorial dispute in the South China Sea. A U.S. Marine Corps assault team moves out of the back of an MV-22 pulling boxes containing a mix of computer chips, printable explosives, and communications gear, and prepares to strike a high-value target. They look more like the cast of MythBusters than Marines. They link up with prepositioned quadcon containers delivered by an unmanned logistics system. The team opens the container and starts assembling a mission payload. After analyzing the different options generated by the Athena computer-assistant, the team leader opts for a mix of hunters and killers: three surveillance drones to find and fix the target, two electronic attack systems to isolate the objective, and three explosive drones trained to target the critical vulnerabilities. One grunt 3D prints explosive charges while another loads new attack profiles for the mission in a tablet using blockly code. They cross check the cloud-based intelligence database and download updates to help the machine-learning algorithm recognize the target and ignore new enemy decoys and civilians. After launching the mission package, the team boards the MV-22 and plans its next attack as it proceeds to a new firing site.
The rapid and creative combination of small, cheap, flexible systems described above represents a new theory of victory: mosaic warfare. The idea emerged in the Defense Advanced Research Project Agency (DARPA), parallel to service concepts like Littoral Operations in a Contested Environment, Multi-Domain Operations, and Multi-Domain Battle. Like these concepts, mosaic warfare describes how to conduct multi-domain maneuver against adversaries possessing precision strike capabilities. Unlike these concepts, mosaic warfare places a premium on seeing battle as an emergent, complex system, and using low-cost unmanned swarming formations alongside other electronic and cyber effects to overwhelm adversaries. The central idea is to be cheap, fast, lethal, flexible, and scalable. Rather than building one expensive, exquisite munition optimized for a particular target, connect small unmanned systems with existing capabilities in creative and continually evolving combinations that take advantage of changing battlefield conditions and emergent vulnerabilities. Put simply, it’s Voltron on the cheap: a human-machine team combining flexible unmanned systems with coup d’oeil (strategic intuition) at a tempo that an adversary cannot match. As forces attack simultaneously from multiple directions they produce a series of dilemmas that cause the enemy system to collapse.
Defense Advanced Research Projects Agency - DARPA. Information itself, when it comes to networking an attacking force to expedite sensor-to-shooter time and quicken the 'kill web,' is increasingly viewed as a defining weapon of war. And now, DARPA, the Pentagon’s secretive Defense Advanced Research Projects Agency, has developed a surveillance tool known as “Memex” that can be used to reveal what is hidden on the dark web, and pursue criminal groups of interest – and perhaps control or even kill the dark web for good. Daniel Javorsek, USAF The goal of the Adapting Cross-Domain Kill-Webs (ACK) program is to provide a decision aid for mission commanders to assist them with rapidly identifying and selecting options for tasking – and retasking – assets within and across organizational boundaries.
Over the past two years, a unique collaboration between DARPA’s Strategic Technology Office, Marine Corps University, and the U.S. Army Reserve 75th Innovation Command resulted in a series of war games to test this concept. This article explores the results. Based on the initial findings, the mosaic concept is a viable way ahead for developing 21st century multi-domain formations and capabilities. The U.S. military should accelerate and support the development of the concept through unified experimentation encompassing people, process, and technology. We need more marines and soldiers, along with coalition partners and scientists, fighting war games and conducting field experiments to transition the mosaic concept into new equipment and tactics that define how America fights.
Mosaic Warfare
Mosaic warfare is the DARPA Strategic Technology Office’s concept for force design and operations to enable 21st century maneuver warfare. Traditional systems are monolithic. From fifth-generation fighters to aircraft carriers, these expensive, exquisite platforms are often static and difficult to adapt. They are systems-of-systems that are like a jigsaw puzzle, where each piece can only interconnect with a matching piece. Alternatively, a “mosaic” system is designed to be flexibly networked and rapidly configured to provide resilient capabilities to the operator. Like the tiles in a mosaic, any system (or unit) having certain functional characteristics could be combined with others to provide a desired warfighting capability at the time and place of a commander’s choosing.
Assembling a task-organized force and its systems for a mission is not new; the Marine Corps does this in the context of the Marine Air-Ground Task Force today. Nor is the idea of using networked systems in warfighting. What is new is the speed and complexity that the mosaic warfare concept allows. The synthesis of flexible force packages with AI-augmented command and control, distributed operations, and mission command provides a powerful basis for modern, information-centric maneuver warfare. A larger diversity of attack options across domains, provided at a speed and scale orders of magnitude faster than today’s platform-centric combined arms or system-of-systems approaches, provides entirely new approaches to preempt, dislocate, or disrupt an adversary. A skilled mosaic commander at the tactical and operational level will employ an ever-changing disposition coupled with the ability to rapidly control and scale signature, combat power, and logistical support to create dilemmas for the adversary, exploit positional advantage, and control tempo.
So, what would mosaic warfare actually look like, and would it really offer any advantage against a capable adversary compared to how we fight today? What happens when we give warfighters the ability to plan and execute all-domain maneuver instead of simply trying to overmatch an adversary using firepower and attrition? While the technological components to realize an experiment at scale are still under development, the results from three recent series of wargames provide a glimpse into the operational potential — and challenges — of translating the mosaic concept into new capabilities and tactics.
Mosaic at Sea
In the first war game series, the Center for Strategic and Budgetary Assessments led two war games in which a notional U.S. Joint Task Force (Blue) fought a hybrid force (Red) with robust precision strike and anti-access, area denial capabilities in East Africa. In this game, a Blue team was tasked with degrading a Red team sensor and weapon complex as part of a larger conflict against Red throughout the Indian Ocean. The games compared the effectiveness of a traditional approach to planning with a mosaic-oriented approach. The mosaic force employed a significant number of unmanned platforms coordinated using a “human command-machine control” scheme in which humans defined strategic objectives and risk, while an AI-supported machine controller translated this guidance into possible options for execution. The war game used a facilitation cell (White) with a mix of subject-matter experts and machine planning aides to replicate this command and control dynamic and dynamically assign resources to missions.
The games illustrated key aspects of a mosaic approach. First, mosaic can be achieved through a composite of existing systems — such as better quadcopter and fires integration at lower echelons — but requires advances in command and control to manage increasingly complex “kill webs.” Unlike traditional kill chains, the webs dynamically source effects from across a network of options based on shifting mission demands. The more intelligence, fires, and logistics assets available to small formations, the more units will need some form of AI-enabled command control and mission planning applications to manage the kill web. This is especially true when small units will be able to fire munitions that can target ground, air, and sea targets. Managing mosaic swarms required human-machine teaming in planning and execution. For example, imagine an AI application that managed resupply requests and personnel replacements, while making recommendations about terrain to the intelligence planner and airspace de-confliction to the fires officer.
The wargames also revealed important insights about risk tolerance and new forms of offensive maneuver among the participants. Players perceived the mosaic force mix as more risk-worthy. They saw losing a low-cost drone in the hunt for gaps to exploit as an acceptable gamble. They viewed cheap unmanned systems as disposable and, rather than put a fifth-generation fighter at risk to probe and suppress defenses, the players employed unmanned systems to assess Red’s disposition and take losses without jeopardizing its main effort. More interesting, however, was how this richer set of options generated different outcomes across the wargames. When Blue players had access to mosaic capabilities, the commander launched a larger number of parallel attacks, as a form of armed reconnaissance, with help from the AI machine controller to identify gaps and exploit them to overwhelm Red’s defenses.
Mosaic in the Combined Arms Fight
The second wargame series leveraged prototype software under development by soldiers from the 75th Innovation Command for Army Futures Command to replicate a tactical fight against a peer competitor with students from Marine Corps University. The scenario involved delaying a Red motorized rifle brigade augmented with additional reconnaissance, artillery, and aviation assets. In the first six games, Blue fielded a U.S. Marine Corps battalion landing team augmented with a small aviation detachment. In the next six games, Blue shifted to a mosaic force with comparable combat power but comprised of 50 percent unmanned systems, including optionally-manned fighting vehicles, small swarms of unmanned aerial vehicles, and long-range loitering munitions with multi-mission payloads. The mission was the same across the 12 games: Blue, outnumbered 5:1, had to delay Red to buy time for additional joint forces to arrive. Each game involved different players on Red and Blue.
In the mosaic treatments, Blue players, similar to the first war games series, employed a swarming form of armed reconnaissance. Different Blue players each leveraged a mix of intelligence, electronic attack, and lethal loitering unmanned systems as skirmishers. These unmanned skirmishers would form a modern version of the old Parthian Shot steppe warfare tactic illustrated in the graphic below. This approach maximized firepower and mobility to wear down an adversary through surrounding them with the equivalent of steppe horse archers. Unmanned hunter-killer teams would find and isolate enemy high-value targets. Like the horse archer of old, these unmanned loitering munitions would circle and attack in periodic pulses designed to degrade enemy capabilities and tempo. The net result was a combined arms defeat-in-detail. Each time, Red lost key systems, and their formation was broken up into smaller, less connected units that could be defeated through fire and maneuver in the deep fight.
Figure 1. Parthian Shot Tactic, adapted from Sean Edwards Swarming and the Future of Warfare
Mosaic in the Electric City
The third wargame series introduced Electro-Magnetic Spectrum Operations effects (e.g. jamming, spoofing, communication relays) and new mosaic planning tools to replicate a futuristic urban fight in a signature-dense environment. The war games took the form of a bracket-style tournament in which teams of 2-5 officers fought 30 one-hour war games. In the games Blue, a light infantry company, fought Red, a motorized rifle company augmented with irregulars, in a race against time to seize key terrain in a city. The players consisted of students at Marine Corps University, supported by soldiers from the 75th Innovation Command. Each side used a prototype synthetic environment, called ULTRA, and an automated planning capability being developed by DARPA’s Prototype Resilient Operations Testbed for Expeditionary Urban Scenarios, or PROTEUS, program. The teams used ULTRA to explore what mosaic warfare looks like for the small tactical unit in a multi-domain urban battlespace.
The battles showed that increasing the number of unmanned systems at the tactical level and integrating signature management platforms tend to overwhelm small command teams. The war games illustrated how different teams dealt with span of control issues associated with unmanned systems and complexity issues associated with signature management. Most players did not have experience visualizing spectrum and thinking in terms of the connectivity of their units, signal strength, and emissions (both friendly and adversary). Even the simplified tool and prototype parameters for visualizing this spectrum in support of maneuver often overwhelmed the majority of participants. Teams that failed to take full advantage of the flexibility of this all-domain mosaic force resorted to conventional tactics that were ineffective and quickly devolved into costly fights for buildings that bled their force.
Teams that succeeded in the signature-rich, dense urban fight maximized speed to shock and dislocate the enemy. These formations tended to maneuver along only two axes to either seize multiple, supporting objectives simultaneously or establish screening or blocking positions in support of the main attack. A contested signature environment forced teams to weight their main effort with additional radio retransmission capabilities to ensure connectivity. Teams used unmanned systems in a reconnaissance and security role that included seeking out and jamming targets of opportunity to control tempo. One particularly effective tactic that emerged across multiple games was identifying and jamming enemy indirect fires assets as a form of counter-battery, freeing up friendly indirect fires to support ground attacks.
In the final matches, the mosaic approach provided useful options for denial and deception to innovative teams. The most successful deception tactic was to replicate the command net of the headquarters in order to limit the ability of the enemy to identify and target central command and control nodes. Yet, employing electromagnetic effects had mixed success. Multiple teams discovered that well-crafted plans including decoys and feints amplified with electromagnetic spectrum operations are reliant on the enemy to actually understand the spectrum and accurately report what they see. That is, you need to understand the adversary to execute even tactical deception effectively, as well as have alternative courses of action ready for when the enemy doesn’t take the bait.
Darpa Kill Web Browser
In this virtual near-future battlespace, the ability to shape signature and perception was critical to survival — and victory. A mosaic force provided a robust and scalable capability to do exactly that, with a risk-worthy force package of multiple unmanned assets that could serve as communications relays, jammers, or strike platforms. But only teams that brought the right skill mix to the match (usually an electronic warfare or communications operator) were able to realize the full potential when combined with conventional combat arms. These teams routinely employed “non-doctrinal” task organizations and tactics throughout the tournament, taking advantage of the additional mobility in the electromagnetic and air domains to scale combat power where needed, rapidly pivoting from defense to offense.
Why Mosaic
Darpa Kill Web Server
The war games reviewed above demonstrate the opportunities and challenges associated with employing a mosaic approach to maneuver. Mosaic warfare builds on long-standing American strengths in combined technological and operational innovation and mission command. Adopting the approach could provide a means to consistently out-innovate and out-offset our competitors. As noted by Bob Work, “the side that finds the better ‘fit’ between technology and operational concepts likely will come out on top.”
But that success will not happen without significant institutional changes. Mission command enabled and accelerated by AI will require shifting from an “industrial age” mindset towards the training and development of warfighters to an information-age approach. Finding, cultivating and retaining personnel — our “Enders” — will be critical to moving from a “salvo exchange” attrition mindset to one that leverages the power of the mosaic warfare concept to maneuver and achieving a position of advantage. More critically, the Pentagon’s machinery for acquisition and budgeting needs to be redesigned for mission-centric acquisition, focusing on concepts rather than gaps, in order to deliver the enabling technology that will make mosaic warfare a reality.
Darpa Kill Web Camera
Benjamin Jensen is a Professor of Strategic Studies at the Marine Corps University School of Advanced Warfighting and Scholar-in-Residence at American University, School of International Service. He is also a Senior Fellow at the Atlantic Council and an officer in the U.S. Army Reserve 75th Innovation Command. The views expressed are his own.
John Paschkewitz is a Program Manager at the Defense Advanced Research Projects Agency, leads the Prototype Resilient Operations Testbed for Expeditionary Urban Scenarios (PROTEUS) program in the Strategic Technology Office, and managed the Center for Strategic and Budgetary Assessment (CSBA) study on Mosaic Warfare. The views expressed are his own.
Image: U.S. Marine Corps (Photo by Lance Cpl. Jacob Wilson)