Military Technology

Swarming, Expendable, Unmanned Aerial Vehicles as a Warfighting Capability

by Gary Martinic

Print PDF

For more information on accessing this file, please visit our help page.

Gary Martinic works as a Team Leader in Technical Support Services at Western Sydney University. In his other role, he is a Flight Lieutenant (AAFC) at 3 Wing Headquarters, Australian Air Force Cadets, located at the Lidcombe Barracks. With a lifelong interest in military aviation and military history, he dedicates a great amount of his time to the supervision, training and mentoring of Air Force Cadets. Gary has a strong interest in unmanned weapons systems of air, land, and sea, and has written numerous articles for national and international defence journals and magazines.


Unmanned Aerial Vehicles (UAVs) have been employed for ISR (intelligence, surveillance & reconnaissance) missions for more than a century, and have been used for strike1 missions2 for more than a decade in modern theatres of conflict as effective light weapons platforms. While these robotic technologies have become relatively ‘mainstream’ today, there has been ongoing research and development (R&D) into the ability of their smaller variants to operate as a ‘single unit’ or swarm,3 with the aim of improving their capabilities and performance with respect to adversary targeting. This ‘evolution’ has occurred primarily because of the tactical advantages that this new developing technology may be able to provide. For example, any military technology that can absorb multiple hits and keep going, from a warfare point of view has a major advantage over other systems, such as manned, and even lone unmanned aircraft, which can be destroyed by a single missile.4 Additionally, the technology can be used in three ways by military forces: to attack, defend and to support functions such as ISR,5 and it reduces the risk of loss of human life and expensive equipment in battle.6

The concept of fusing UAVs into swarms has seen two key developments stand out in particular:

  • the ability to swarm shortly after being launched from either a pneumatic catapult from an aircraft, a ship or from a submarine; and
  • making them inexpensive enough so as to make them ‘expendable.’

Photo courtesy of US Office of Naval Research

Figure 1 – Tube launching of a Raytheon Coyote small UAV.7


With Adolf Hitler's adoption of advanced tactics and technology, such as advances in communications through Blitzkrieg warfare, and weapons, such as the jet-powered Me-262 fighter and the V-1 ‘buzz bomb’ and V-2 ballistic missile rockets fielded in the latter stages of the Second World War, the Allies were finding themselves constantly ‘behind the curve’ in the technology of war. And this changed the way that they planned to fight in the future.8 Over the next 70 years, western military powers have sought to lead the way in aerospace and weapons research and development (R&D). Consequently, the US and other western powers made sure that they held a clear advantage with respect to tactics and technology, often a generation ahead of potential adversaries, by replacing the focus from quantity to quality, so that they could deter any adversary.9 Today, this race to maintain military supremacy has been extended into R&D pertaining to unmanned weapons systems (UWS) with air,10,11,12 land,13,14,15 and sea applications.16,17,18 Indeed, the world’s most advanced militaries continue to develop Unmanned Weapons Systems because of the significant tactical advantages that they provide. While these robotic technologies have advanced significantly across all three environments, they arguably have been the most pronounced in the air, with unmanned combat aerial vehicles (UCAVs).19 The logical extension of this technology is its applicability to UAV swarming, where, just as in nature, swarming systems have individual agents that interact with one another and their environment. These agents follow simple rules, but the collective interactions between the agents can lead to quite complicated and sophisticated collective behaviours, including emergent behaviours, and even intelligence aspects. For example, a swarm may stay in formation while changing direction several times.20 In order for this to be achieved, individual units must be physically-homogenous with the same programming and the same sensors, which enables an autonomous swarm to communicate with each other, noting that the sensors are used to disguise swarm behaviour, which are often based upon environmental factors outside the swarm.21

Sueddeutsche Zeitung Photo/Alamy Stock Photo CPMPHP

A German V-1 rocket prior to launch.

INTERFOTO/Alamy Stock Photo B3EH0J

German V-2 rocket on a Mielier vehicle, circa 1943.

Currently, medium-size UAVs are optimised for ISR and light strike operations in non-contested or relatively permissive environments. However, significant advances have been made in developing the next generation of smaller UAVs with the ability to swarm,22 in order to attack specific military targets.23 This has the added advantage that they waste enemy resources by drawing fire,24 or alternatively, they could be equipped to jam enemy communications via on-board sensors. This generation of small UAVs has been developed to be modular, adaptable and inexpensive, given that the payload they carry determines the type of mission they can execute. Such small UAVs have been found to be a cheaper and more cost-effective all-round military technology when one compares the costs to, for example, the F35 Joint Strike Fighter programme, which has cost approximately 1.5 trillion US dollars, to date. With most naval anti-ship and air to ground missiles costing upwards of a million dollars each, the goal has been to cost-effectively produce an entire swarm of small UAVs costing less than a single missile. This goal has already been achieved with Raytheon’s Coyote small folding wing UAV (see Figures 1 and 2), which cost around $15,000 USD per unit,25 with the challenge being to reduce the costs even further to somewhere between $5,000-$7,000 USD per unit.26 Indeed, UAVs of all classes have taken on offensive capabilities with the integration of adapted and purpose-built munitions, and look set to take on more roles as their capabilities are expanded and improved. For example, Defence Advanced Research Projects Agency (DARPA) are currently working on armed ‘deploy and recover’ UAVs, which can also be launched from a ‘mothership,’ as shown at Figure 3, and which are recovered post-mission.27

Redrawn figure from Swarms of drones ‘willthink for themselves’ by Deborah Haynes.

Figure 2 – Cheap and expendable armed UAVs being tube-launched.28

Click to enlarge image

A significant milestone in the future of air warfare was achieved in late-2016, when the US Navy successfully demonstrated that a flight of around 30 Coyote UAVs could be fused into operating as a single swarm, above the ocean, and at an undisclosed location.29 The mission was intended to show that the swarm could be self-configuring, so that if one UAV was destroyed, others in the swarm could autonomously30 change their behaviour and complete the mission. Thus, small UAV swarm systems, which are aware of each other’s position and movements, have been an incredible advance, meaning that UAV swarms can be much harder to stop.

Redrawn figure of DARPA Gremlins from

Figure 3 – Cheap and expendable armed UAVs being launched from an aircraft.31

In these tests, the UAVs also demonstrated that they could position themselves autonomously, flying in formation without being directed where to go, which, as opposed to remotely controlled operation, represents a major evolutionary leap forward, since the swarm effectively displayed ‘collaborative behaviour.’32 The Coyote UAVs are a metre-long tube-launched, electrically-powered small UAV. Designed to be an expendable asset used for reconnaissance, this UAV has folding wings, so it can be fired from the tubes used for dropping sonobuoys on anti-submarine aircraft, or from a pneumatic launcher on a navy ship. Weighing around 6 kilograms, once launched, the Coyote’s wings flick out and it can fly for up to 1.5 hours on battery power, while at the same time beaming back video messages from 30 kilometres away.33 Coyotes were also used by the US Office of Naval Research (ONR) in a programme known as ‘Low-Cost UAV Swarming Technology’ (LOCUST), which was designed to demonstrate whether autonomous, swarming small UAVs can overwhelm an adversary more cost-effectively than conventional weapons systems.34

The impressive thing about the LOCUST testing by ONR is that they launched 30 UAVs within 40 seconds, upon which the UAVs rapidly formed into a swarm, and then flew autonomously in formation to carry out the mission, communicating by using a low-power radio–frequency network, which enabled position sharing and other data.35 As endurance is limited to 90 minutes of operation, rapid launch was crucial for the battery-powered UAVs, which were designed to be platform, payload and mission-agonistic.1,15 The swarming mechanism used was a ‘parent/child’ relationship, in which one of the UAVs acts as the lead, and the other UAVs follow. However, the ‘leader’ can also be changed in case it is destroyed during the mission.1 Interestingly, using certain electronic commands, the operator can redirect individual UAVs to perform other missions, and the swarm can also be broken up into smaller groups for alternative manoeuvres, or a single UAV might break formation36 to get a closer look at a target, and then return to carry out an attack.37 These scenarios indicate that a significant degree of formation control has been achieved, along with other vital data collected, which included how tight the formation could fly as a swarm, at what altitude, and what type of manoeuvres it could perform.38

In October 2016, the USN also successfully launched 103 miniature swarming drones from F/A-18 fighter jets, which was carried out at an undisclosed location.39 Then, in early-2017, the USN carried out similar tests at the Naval Air Weapons Station China Lake Test Range Facility in California.40 In both tests, Perdix micro-UAVs successfully demonstrated advanced swarming behaviours, such as ‘…collective decision making and adaptive formation flying.’ These Perdix low-altitude micro-drones were not pre-programmed, synchronised individual units. They were a collective ‘organism,’ sharing one distributed brain for decision-making and adapting to each other like swarms in nature.41 As every Perdix communicates and collaborates with every other Perdix, the swarm has no leader, and can gracefully adapt to drones entering or exiting the team.42

Previous successful demonstrations have included an airdrop from F-16 fighter jet flare canisters by the US Air Force Test Pilot School at Edwards Air Force Base in 2014.43 The US Navy have also successfully launched X-wing-shaped small drones vertically into the air, after being fired from the torpedo launch tube of a submerged US submarine, the USS Providence, in December 2013.44

DVIDS 747118/US Navy photograph by Lieutenant (JG) Jeffrey Prunera

The Los Angeles Class attack submarine USS Providence transits the Thames River as it departs Naval Submarine Base New London for a regularlyscheduled underway.

While these demonstrations by ONR and others have been impressive, there are still hurdles needing to be overcome before these new capabilities become fully established. Firstly, autonomous ‘sense-and-avoid’ technologies in small UAVs are still in their early developmental stages, and solutions will need to be found, although as processors are getting more powerful and reliable, this issue is likely to be resolved via the use of deep learning and neural networks as technology advances.45 This is important, because it’s one thing to fly a swarm above open water, but then it’s quite something else when that swarm needs to be flown above land where there are numerous obstacles to avoid, such as buildings, power lines and trees, let alone a land warfare scenario, where there may be adversary weaponry with which to contend.46

Secondly, there are two other issues in establishing trust with respect to completely- autonomous systems, which again, are likely to be overcome eventually as more tests are safely completed. Thirdly, there is the issue that a swarm’s endurance is limited by the duration of its battery life. Again, a potential solution to this problem is to establish a ‘hive’ or base station, where individual UAVs can return for recharging while the rest continue with their mission.47

Lastly, public safety policies predominantly treat unmanned aircraft as if they are manned, meaning that they are highly regulated if they endanger public safety, or enter civilian airspace. The issue here being that policy makers will be more cautious as they are dealing with UAVs being operated fully autonomously, as opposed to being remotely piloted, which is still preferable from a flying safety standpoint.48 It is important to compare the differences between the two systems at this juncture. An RPA is the acronym for a Remotely Piloted Aircraft, which is a form of an unmanned aerial system (earlier acronyms for this were UAS), which is non-autonomous in its capacities, the aircraft being subject to direct pilot control at all stages of flight despite operating remotely from that pilot.49 Swarming UAVs are flocks or groups of small UAVs that can move and act as a group with only limited (semi-autonomous) or no (autonomous) human intervention.50 These systems also differ in that RPAs usually have a much longer flight duration (or loiter times), whereas swarming UAVs, currently have a limited flight duration of up to 1.5 hours maximum (although R&D continues into methods which may keep them in the air for longer periods). Lastly, RPAs should be considered as a safety critical system, as they often fly in and out of civilian airspace. Some authorities consider the risks posed by swarming UAVs as being too great, and advise that those risks should be considered sooner rather than later before their destructive potential reaches maturity.51 Swarming UAVs can be considered both safety critical and mission critical systems, although they are primarily a mission critical system (as indeed are weapons), and it is for this reason that they should not be released into civilian airspace other than for the purpose of an authorised military mission.52 In this interesting new source, future human decision-making regarding complex military and safety critical systems is analysed in detail. It addresses the likely changes to weapons, cyber warfare and artificial intelligence (intelligent and autonomous systems) to emphasize that these new capabilities need to be thoroughly tested before being fielded, in order to ensure they are safe and operationally effective, while at the same time, mitigating unintended hazards and adverse effects. They also provide detailed explanations with respect to meaningful human control53 during the Find, Fix, Track, Target, Engage, Assess (F2T2EA) ’kill chain,’ with recommendations for ethically-aligned design of AI and autonomous weapons.

The advantage of robotic swarms is that they are not limited to one particular military domain only,54 as they will likely prove be equally effective across all domains, particularly when used in combination with the advanced tactics that they were designed to undertake, whether they may be offensive or defensive in nature. To this end, drone swarms could be used to blanket enemy areas with ISR assets, to jam enemy air defences, and to overwhelm enemy targets with firepower. They would likely be particularly useful in all phases of the F2T2EA targeting cycle, and as an alternative to precision-guided munitions (PGMs), because even the most advanced PGMs become useless if targets cannot be located and designated for attack.55

UAV swarming technologies and tactics bring significant changes to warfighting capabilities, including the ability for ‘kamikaze-style’ attacks to overwhelm adversarial assets, which can include neutralising enemy missile batteries, radar stations and other systems, or by rendering same sites vulnerable to attacks by more heavily-armed manned aircraft. They can also be deployed to conduct important ISR and other imagery missions deemed too heavily defended to be carried out by manned aircraft. UAV swarms can also be employed in defensive roles where they can protect larger navy ships, heavy armour and artillery, or large aircraft assets from attack by establishing defensive barriers. Indeed, LOCUST is part of an effort to develop autonomous technologies that can be applied across surface, undersea and air domains.56


The growing importance of unmanned vehicles stands as a testament to the evolution of military technology.57 It is the author’s view that UWS, including swarming UAVs, are the future of warfare. The ISR-gathering value of unmanned vehicles is well-demonstrated, as UAVs can remain on station over areas of interest sometimes for days at a time, making them one of the most valuable persistent-surveillance platforms available.58 As a weapons platform, UAVs with light missile armaments have taken out attacking forces, and have killed many terrorist leaders in the Middle East in precision surgical strikes.59 Furthermore, unmanned vehicles help keep humans out of harm’s way. As a result, battlefield casualties can be reduced, and UAVs cut down on the possibility that a human aircraft pilot will be shot down, taken captive and remain in the headlines for months, if not years.60

Research into swarming UAVs is one of the fastest and most promising areas of military R&D today, as swarms have essentially advanced the capabilities of UAVs even further. These has been possible mostly because of algorithms, in that their application is that which governs swarm behaviour, making communication and cooperation possible within the swarm. Essentially, UAV swarms are low-technology hardware knitted together with high-technology artificial intelligence (AI).61 This combination will likely become a powerful weapon of the future, including both lethal and non-lethal applications, enabling essentially a light attack force to defeat more powerful and sophisticated opponents.62 Such algorithms will enable UAVs and UAV swarms to conduct a much wider range of functions without needing human intervention, such as sensing, targeting, weapons adjustments and sensor payloads, range and capabilities.63 Developments with respect to AI will better enable unmanned platforms to organise, interpret and integrate functions independently, such as ISR filtering, sensor manipulation, manoeuvring and navigation; hence emerging computer technology will better enable UAVs to make more decisions and to perform more functions by themselves.64 The advent of swarm technology heralds a period that could reverse the trend of the past quarter of a century of US military dominance, which has seen the deployment of fewer but more advanced – and expensive- weapons platforms. The next generation of weapons may see sophisticated technology systems outdone by the sheer numbers of autonomous swarms.65

Just when the US achieves its goal in developing these new UAV swarming capabilities ready for acquisition and deployment as front-line weapons systems remains to be seen. The results so far have demonstrated that it is well on track to meeting research goals in the near-future, as it continues development and testing on a range of systems and levels of autonomy. The creativity and innovation of these projects represents an unprecedented paradigm shift in small UAV launch systems, strategy and tactics with the myriad modes of operation, and the technology certainly has the characteristics to be a ‘game-changer’ for the US and its Allies.

DVIDS 5356101//US Army photo by Pv2 James Newsome

The US Army 11th Armored Cavalry Regiment and the Threat Systems Management Office operate a swarm of 40 drones to test the rotational unit capabilities during the battle of Razish, National Training Center on May 8th, 2019. This exercise was the first of many held at the National Training Center.


  1. The ability to conduct a military bombing mission with pinpoint precision and efficacy. As per G. Martinic, ‘Drones’ or ‘Smart’ Unmanned Aerial Vehicles? in Australian Defence Force Journal. Issue 189, Nov/Dec 2012, p.1.
  2. For an example of a precision surgical strike mission, see Associated Press, ‘Abu Yahya al-Libi, al Qaeda deputy leader, killed in U.S. drone strike.’ CBS News, 5 June 2012, at:
  3. To operate as a single unit. As per D. Hambling, ‘Drone swarms will change the face of modern warfare,’ in The Wired World , 2016. p.1
  4. Ibid.
  5. P. Scharre, Robotics on the Battlefield, Part II: The Coming Swarm. Washington, D.C: Center for a New American Security, 2014.
  6. L. Lachow, ‘The upside and downside of swarming drones,’ in Bulletin of the Atomic Scientists.,73 (2), 2017, pp. 96-101, p.96 and p.98
  7. Photo courtesy of US Office of Naval Research (sourced from Accessed 16 November 2017.
  8. J. Ostberg, ‘Warfare has changed forever now that there are no secrets,’ in WIRED, December 2015.
  9. Ibid.
  10. Martinic, ‘Drones’ or ‘Smart’ Unmanned Aerial Vehicles?
  11. Senate Report, Use of unmanned air, maritime and land platforms by the Australian Defence Force, 2015, ISBN 978-1-76010-227-2. Australian Government Press, pp. 11, 22-23, and 45, for invited contributions by the author.
  12. Editorial. ‘Unmanned aerial warfare. Flight of the drones. Why the future of air power belongs to unmanned systems,’ in The Economist. 2011, p.5. Also online at:
  13. G. Martinic, ‘The proliferation, diversity and utility of ground-based robotic technologies,’ in Canadian Military Journal. Vol. 14, No.4, 2014.
  14. E. Sofge, ‘America’s robot army: are unmanned fighters ready for combat?’ in Popular Mechanics, January 2013, p.2. Also available online at:
  15. US Committee on Army Ground Vehicle Technology (2002). Technology development for Army ground vehicles. Pp.1-2, 5. National Research Council, National Academies Press, Washington.
  16. G. Martinic, ‘Unmanned maritime surveillance and weapons systems,’ in Headmark, Journal of the Australian Naval Institute. Issue 151, March 2014.
  17. US National Academy of Sciences, Engineering and Medicine, ‘Mainstreaming unmanned undersea vehicles into future US naval operations,’ in National Academies Press, 2016.
  18. E.C.Whitman, ‘Unmanned underwater vehicles: beneath the wave of the future,’ in Undersea Warfare Magazine, September 2013. Whitman reports that the LOCUST project manager had been working closely with the Georgia Technological Research Institute to develop the system in which individual UAVs could position themselves autonomously, while flying in formation without being directed explicitly where to go.
  19. Martinic, ‘Drones’ or ‘Smart’ Unmanned Aerial Vehicles?
  20. Lachow.
  21. Ibid.
  22. Aggregation refers to the UAV’s ability to join the swarm, as per Hambling, p.2
  23. Ibid, p.1.
  24. Ibid.
  25. N. Lee, Video: ‘Watch the (US) Navy’s LOCUST launcher fire a swarm of drones,’ at Business Insider Australia, 2017.
  26. G. Warwick, ‘ONR: Swarming UAVs could overwhelm defences cost-effectively,’ in Aviation Week & Space Technology, pp.1 and 2, 2015.
  27. B. Stevenson, ‘DARPA seeks information on manned UAV mothership,’ in, November 2015, p.1.
  28. Redrawn figure from Deborah Haynes Friday, Swarms of drones ‘will think for themselves’, in The Times & Sunday Times, 17 April 2015. Also sourced from Accessed 16 November 2017.
  29. Hambling, p.1.
  30. Autonomous here indicates that once launched, the UAVs will perform a mission that was pre-programmed by humans. The swarm is not self-aware, and does not make decisions of its own (apart from navigation to stay on mission, etc).
  31. Redrawn figure of DARPA Gremlins from Accessed 19 November 2017.
  32. Hambling, p.1
  33. Note that Tenix had proposed the use of the ACR Coyote for release from AP-3C Orion aircraft sonobuoy tubes for an unsuccessful Capability Technology Demonstrator option in 2008 for eventual use with the P-8 Poseidon.
  34. Lee.
  35. Ibid.
  36. Disaggregation refers to the UAV’s ability to leave the swarm, as per Hambling, p. 2
  37. Hambling, p .2, and Warwick, p.2.
  38. Ibid.
  39. C. Baraniuk, ‘US military tests swarm of mini-drones launched from jets,in Technology, US Department of Defense. 2017, p. 1.
  40. S. Snow, ‘Pentagon successfully tests world’s largest micro-drone swarm,’ in Pentagon & Congress, 2017, p.1.
  41. Ibid.
  42. Ibid.
  43. Ibid.
  44. T. Eshel, ‘X-wing drone launched from a submerged submarine for the first time,’ in Defense Update, December 2013.
  45. Hambling, p.3.
  46. Ibid, and Scharre, p. 31.
  47. Hambling, p.3.
  48. Ibid, p. 2.
  49. As per Australian Certified Unmanned Aerial Vehicle Operators, ACUO. Sourced 18 March 2018.
  50. Lachow, p.1.
  51. Lachow, among others …
  52. M.G.Tutty and T. White, ‘Unlocking the future: decision making in complex military and safety critical systems,’ Systems Engineering Test and Evaluation Conference Sydney 30 April – 2 May 2018.
  53. IEEE EADS V2.0, Aligned Design: A Vision for Prioritizing Human Well-being with Autonomous and Intelligent Systems, Version 2. IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems, ISBN 978-0-7381-xxxx-x, 3 Park Avenue, New York, NY 10016-5997, USA, December 2017. [Online, accessed 15 December 2017].
  54. The domains of war are as follows: physical (including land [and sub-surface], air, space, and sea [and sub-surface]), human (cognitive and social), and information, or PHI.
  55. D. Turnbull, Is relying on smart weapons a smart approach? in Australian Defence Force Journal. Issue 204, 2018, p.37.
  56. Warwick.
  57. J. Keller, ‘Military and aerospace electronics gives unmanned vehicle technology the attention it deserves,’ in, December 2013.
  58. Martinic. ‘Drones’…, Senate Report, and Editorial.
  59. Martinic, ‘Drones’…
  60. Keller, p. 2
  61. E. Feng and C. Clover, ‘Drones swarms vs conventional arms: China’s military debate, in Financial Times, 2017, pp. 2-4. See also:
  62. Ibid.
  63. J.K.Osborn, ‘Drones of tomorrow will be smarter, stealthier, and deadlier,’ in Warrior Scout, 2016, p. 1. See also:
  64. Ibid.
  65. Feng and Clover.