Estimated time to read: 8 minutes
By Caitlin Thorn
Excerpt: Delivery of swarming systems to the warfighter will not be hindered by the technology development, but rather the processes and procedures in which to implement this technology.
A previous Drop Zone article here at OTH discussed the Third Offset Strategy and its emphasis on the application of autonomous systems. Autonomous capabilities are not a new technology in itself, but the application of autonomous technologies in the military sector has yet to be fully exploited. Currently, autonomous technology is primarily used in unmanned systems, successfully facilitating combat and intelligence, surveillance, and reconnaissance (ISR) missions without the risk to a pilot. As autonomy negates the need for manned aircraft, this opens the door for applications that would otherwise have been deemed too expensive or risky through the use of manned vehicles. One such example is the application of autonomy to swarming systems. Swarming systems hold incredible potential for offensive capabilities as they are flexible, adaptable, and challenging to defend. The technology to enable this capability is surprisingly simple—autonomous systems, intra-swarm datalinks, and cyber-enabled algorithms have existed for a decade or more. It is the overdue development and tailoring of the tactics, techniques, and procedures (TTPs), acquisition processes, and DOD policy for the application of these systems that preclude operational use. The focus of this article will be the implications of the development of swarming weapons systems for use in the future fight, to include tailoring the design, processes, and policies to the capability that swarming weapons demand in the future operating environment. The development and use of algorithms to enable the TTPs of the swarms themselves also hold significant implications for the effectiveness of these systems, and will be addressed in a follow-on article.
Swarms are unique compared to any other offensive weapon the Air Force has developed or employed in that its capability is derived from the functionality of a collective group of weapons, and not the individual weapon itself. This idea holds several implications for the design of the individual weapon, the most significant of which is affordability. Typically, weapons must be designed to perform a lethal mission with a high reliability. The capabilities and testing required to ensure high performance and reliability for each unit drives a tremendous amount of cost. A swarming weapon system designed for a specific mission will be comprised of multitudes of weapons—the exact number will vary according to the scenario and mission set, but could be anywhere between tens to hundreds of weapons. As the capability is a result of the collective performance of the group, it is possible to design individual weapons with less functionality and reliability without sacrificing the performance of the swarm. This allows for affordable individual weapon units. It is probably safe to say that affordability is not generally a priority for a weapon, but for a swarming system it may be a design constraint. This results in designing individual weapon units to meet a price ceiling, which leads to a rather simple and relatively low threat weapon individually. However, the system threat increases exponentially when the collective capability is exploited by a swarm driven by scenario specific TTPs. Affordability also allows for the weapon to be expendable, which in turn eases the development of TTPs to allow for mission success with an acceptable individual weapon loss ratio. And finally, less capable and affordable weapons generally result in a simpler individual weapon design, allowing for a modular/open architecture approach. This allows for multi-variant weapons that may fulfill differing capability needs and perform unique mission sets for minimal investment.
Currently, the path to develop and deliver technologies to the warfighter is through the standard DOD acquisition processes. Historically, it has been clearly demonstrated that product delivery to the warfighter is a timely manner is severely hampered by the bureaucratic procedures that these processes entail. Yet, despite the delay between tech development and product delivery to the field, the Air Force has managed to maintain technological superiority. As the speed of technological advancement continues to increase at an alarming pace, the Air Force may no longer afford this luxury. Faster delivery to the warfighter will be necessary to field technologies before they become obsolete. Developing and fielding swarming weapons systems are no exception, and fortunately their unique capabilities may facilitate tailoring these processes to accommodate a compressed development and delivery schedule. As mentioned previously, a less capable, affordable design allows for simpler, modular systems. This modularity allows for the development and delivery of individual weapon variants at a fraction of cost and schedule. This type of spiral development process allows for significant schedule savings, as the modular design allows for components to be added or swapped without a complete system redesign. This enables a flexible, iterative development process, enabling a faster transition from tech development to program acquisition.
Also previously mentioned, as the capability of the system is derived collectively from the swarm, the TTPs may be tailored to allow for a degree of individual weapon attrition. This allows for less reliable weapons and thus less rigorous testing processes. As current weapons in inventory are reliant on an individual unit’s performance, individual weapon reliability must be extremely high. Testing processes to ensure this reliability contribute to additional expenses and prolonged schedules. Designing for a relatively less reliable individual weapon while still allowing for a high collective swarm reliability may be a key component to decreasing schedule and further lowering costs. Also, as the individual weapons are rather simplistic themselves, individual testing should not be schedule intensive.
But perhaps the most challenging process implication for swarming systems will be the verification and validation of the swarm behavior. Swarm behavior will be controlled through the EMS domain, with algorithms executing swarm tactics. No code is 100% reliable, and there is no DOD guidance or procedures for risk tolerance for validation and verification efforts for semi-autonomous swarm behavior. In addition, although most of the swarm behavior testing will take place via modeling and simulation efforts, physical tests on Air Force ranges are still necessary. Although perhaps limited in scope, Air Force ranges are relatively unfamiliar and unprepared for multi-agent tests, and must start preparing now in anticipation of future requirements.
Established in 2012, DOD 3000.09 is the DOD’s first attempt to establish guidance on the use of semi-autonomous and autonomous weapon systems. As the US makes the leap into uncharted territory with the use of autonomous capabilities in weapon systems, it seeks to comply with international humanitarian law that requires a responsibility to distinguish civilians from combatants. To address the concern that autonomous systems lack the human cognitive capability to make this distinction, the directive clearly prohibits the use of lethal force for autonomous weapon systems. It does, however, allow for the use of lethal force for semi-autonomous weapon systems. Although this is clearly stated, the vagueness of the directive lies in how autonomous and semi-autonomous systems are defined. According to the directive, an autonomous weapon system is defined as one that can select and engage targets without further intervention of a human operator. A semi-autonomous weapon system is defined as one that is intended to only engage individual targets or specific target groups that have been selected by a human operator. The definition goes on to state that “fire-and-forget” homing munitions are also considered semi-autonomous weapons systems if they rely on TTPs to maximize the probability that the only targets within the seeker’s acquisition basket that are those that have been selected by a human operator. As a “fire-and-forget” system uses a seeker to autonomously select targets, this then translates to the munition becoming, de facto, autonomous. This degree of ambiguity between the two definitions muddles the line between the differences of a semi-autonomous and autonomous system, and opens the door for these lines to be crossed in the implementation of this technology and likely contribute to further delays in delivering this technology to the warfighter.
Of the possible next generation “game-changing” technologies, autonomous capabilities may have the most significant implications for defining the future fight. The applications for autonomy are numerous, from the current use of this capability to facilitate the operation of unmanned vehicles, to the future capability to replicate human cognitive ability and decision making processes. But perhaps the most intriguing application of autonomy lies in the application of swarming systems. The use of this capability holds tremendous potential as an offensive weapon force multiplier, as its inherent flexibility may be very hard for adversaries to defend against. As our adversaries are closing the gap on the potential of this technology, failing to take into consideration the design, processes, and policy to fully exploit the potential of swarming weapon systems in a relevant timeframe will most likely result in the failure to provide the “offset” critical to maintaining our technological advantage in the future operating environment.
Caitlin Thorn is an aerospace engineer in the United States Air Force. She is currently attending Air Command and Staff College at Maxwell AFB, AL.
The views expressed are those of the author and do not necessarily reflect the official policy or position of the Department of the Air Force or the US government.