19 Mar Drone Swarming and the Explosive Remnants of War
[Maziar Homayounnejad is currently a PhD researcher at the Dickson Poon School of Law, King’s College London. His research primarily focuses on law of armed conflict aspects of autonomous weapon systems, with a secondary focus on arms control and non-proliferation.]
On January 5th of this year, a Russian air base and a nearby naval base were attacked by a swarm of 13 makeshift drones carrying explosives. This was the first known swarm attack to take place in a real battlefield and, fortunately for Russian forces, it failed: according to the Ministry of Defence, seven were shot down, while the remaining six were electronically intercepted. That said, the audacious attack has revived talks of drone swarming becoming both a future terrorist threat and a new style of warfare for traditional armed forces. Inevitably, any trend that sees more munitions deployed in the same space will raise questions on the possible humanitarian effects.
Drone Swarming in Context
Despite all the attention focused on the January attack, a serious discussion had already emerged on the use of battlefield drone swarming. Arguably, the current debate began in a seminal report by Arquilla and Ronfeldt in 2000, which pointed out that a) swarming will be an evolution of the centuries-old military doctrine of mass and maneuver, b) this seemingly amorphous tactic will in fact be “a deliberately structured, coordinated, strategic way to strike from all directions”, and c) advances in sensor and communications technologies will provide the requisite connectivity to make all this possible (as will be seen below, these principles have all stood the test of time). Accordingly, it was thought that drone swarming will become an irresistible option for both professional militaries and non-State actors alike, as commercial sector electronics advance and miniaturize.
Building on this, in a 2014 report entitled ‘The Coming Swarm’, Paul Scharre contended that drone swarming will provide a viable solution to many of the challenges faced by the US military. The problem, he argued, is that US defense capabilities may weaken in relative terms through a combination of a) the improved long-distance precision targeting capabilities of near-peer adversaries, and b) exponential rises in the cost of large military aircraft and ships, while US defense budgets only grow linearly (one of “Augustine’s Laws”). This would leave a dwindling attack capability vulnerable to easy defeat by a more accurate, precise and numerous adversary. On the other hand, by reducing the size and increasing the quantity of combat assets, US forces will benefit from “Lanchester’s Square Law.” This asserts that “twice as many units in the fight actually translates to a fourfold increase in combat power for units with aimed-fire weapons”, because while US forces can ‘double up’ on attacking enemy units, the other side can only attack half the US forces at any one time (see summary in Report Preview). The implication was clear: swarming micro-drones will become a game-changer, and will be financially viable in the face of (then) defense sequestration.
Subsequently, and from a slightly different angle, David Hambling argued in Swarm Troopers that the abundance of cheap, powerful and rapidly-advancing dual-use components from the mobile phone industry will make micro-drones and swarming capabilities easily available to non-State actors. The stark reality, Hambling argued, is that States will have to choose between a $100m fighter jet, which will be vulnerable to a single enemy strike, or 50,000 micro-drones, which will offer unparalleled resilience and attack capability when facing enemy fire. This is because “if one drone gets taken out, the others autonomously change their behavior to complete the mission.” Namely, while a single enemy strike can down the entire $100m jet, the same will usually only destroy one (or a few) micro-drones, while the overall swarm system remains intact and can carry on in battle. Again, we see suggestions that drone swarming will become an irresistible option for professional militaries; partly because of the options open to non-State actors, but with the added result that the US and near-peer adversaries like China will enter a swarm race, with potentially strategic consequences.
Pinning Down the True Nature of Drone Swarming
To be sure, the January attack on the two Russian bases was not a ‘true’ example of swarming, as each of the 13 drones appeared to be operating independently, albeit via GPS and in close proximity. As Scharre points out, true swarms will not only bring “greater mass”, but also “coordination, intelligence and speed to the battlefield” for smart maneuver (pages 5, 10 and 12 of The Coming Swarm). This suggests both a quantitative (sheer numbers) and a qualitative (collaborative algorithms) dimension. Accordingly, the most effective drone swarms will not only overwhelm and saturate enemy defenses, but will also outsmart them with “cooperative behavior among distributed elements that give rise to a coherent, intelligent whole” (page 10). Examples include autonomous regrouping after an attack, as mentioned above; or confusion tactics, such as micro-drones autonomously disaggregating to avoid detection, before reaggregating at the last moment to take the enemy by surprise.
The January attack displayed no such element, but we may expect that weapons designers will study how and why this ‘Model-T’ swarm was defeated in order to inform the development of their own collaborative algorithms for future deployment. At the very least, it is possible that the January attack inspired another swarm attack the following month by Houthi forces against the United Arab Emirates. While the status of this most recent attack is unclear, it does appear to have been more carefully executed, and it may indicate a new trend in armed conflict, against which US ground forces are currently ill-equipped to defend.
So far, only explosive micro-drones have been discussed. In reality, swarms may be developed to undertake a variety of roles, including intelligence, surveillance and reconnaissance and/or electronic warfare. However, it is swarming for kinetic attack that will pose the greatest potential humanitarian risk; those designed for offensive land attack (as opposed to naval defenses), even more so.
The ERW Aspect of Drone Swarming, and Ways to Mitigate It
Ultimately, the aim of any future drone swarm is to outsmart, saturate and overwhelm the enemy, such that even a sustained defensive attack will not prevent a few ‘leakers’ from getting through to kill, destroy or neutralize their target. This leads to two distinct but related humanitarian problems.
Firstly, where a large number of non-leakers are inevitably shot down, they will remain on former battlefields; should they fail to explode as intended, these micro-drones will very likely become explosive remnants of war (ERW), giving rise to post-conflict civilian risks. Potentially aggravating the situation is the possibility of an ‘arms race’, as noted above. As Lachow explains (at page 100), this may see ever-increasing swarm sizes being met with stronger defensive systems and increasingly large counter-swarms, leading to yet greater offensive swarms, more non-leakers being shot down, and so on. The result would arguably be a heightening of the ERW problem.
Secondly, there is the problem of the physical appearance of the micro-drones, which can bear a striking resemblance to toy airplanes. As noted in an earlier Article 36 report (at pages 10 and 44), the size, shape and color of munition remnants can make them attractive to children, who unwittingly pick them up in the expectation of playing with a toy. Downed micro-drones, particularly those that are brightly-colored, and of approximately the same size and appearance as toy airplanes, will be highly alluring to children, who may find that the apparent toy leaves them with life-changing injury.
Arguably, these post-conflict risks merit closer attention before drone swarming is deployed in significant numbers.
In a recent paper published by the Transnational Law Institute at King’s College London, I examine the drone swarming ERW problem, and provide a legal commentary of the potential application of three existing instruments that may help to address it: Protocol V of the Convention on Certain Conventional Weapons, the Convention on Cluster Munitions, and Amended Protocol II.
Protocol V (pages 22-47 of the paper) specifically addresses ERW, and it creates a number of obligations for States and non-State parties alike, though these are heavily caveated. For example:
- Article 3, the centerpiece of the Protocol, requires parties in control of territory to mark and clear, remove or destroy ERW as soon as feasible after the cessation of hostilities. Parties not in control of territory shall, where feasible, provide technical, financial, material or human resource assistance to facilitate the marking and clearing, removal or destruction of ERW.
- Article 4 requires that before and during the deployment of munitions that may become ERW, parties shall to the maximum extent possible and as far as practicable record and retain information on the use of such munitions, to facilitate the discharge of Article 3 obligations, amongst others. After the use of such munitions, parties shall, without delay after the cessation of active hostilities but as far as practicable, subject to their legitimate security interests, make available relevant information to the party or parties in control of the ERW-affected area.
- Article 9 encourages States to take ‘generic preventive measures’ aimed at increasing munitions reliability, to maximize the chance that they will explode as intended, thereby minimizing the ERW risk. Such measures are largely based on the Technical Annex to the Protocol, which itself contains voluntary measures.
There are numerous other measures, the provisions of which are impressively comprehensive; yet, it is recognized that these may risk becoming both labor-intensive and financially burdensome, if mandatory. Consequently, virtually all Protocol V provisions are heavily caveated, as indicated by the italics in the above bullet points. Sadly, this may provide a ‘get out clause’ for States and other parties that do not wish to act in good faith. On the other hand, such caveats may be interpreted restrictively in the case of swarming munitions, in light of their expected technical features. Unlike ‘dumb’ munitions, micro-drones are likely to possess a range of ‘smart’ features, such as GPS, advanced electro-optical/infrared sensors, radio communication links, and (potentially) integrated diagnostics (see pages 15-19 of the paper). Collectively, these will do much of the ‘heavy lifting’ of data recording, retention and transmission, such that they remove a significant weak link, which has so far discouraged States from applying Protocol V more expansively.
Of course, some of these technical features will add to production costs, thus they may need to be legally mandated. For this reason, there are possible transplants that may be entertained from the Convention on Cluster Munitions (see the Article 2(2)(c) technical criteria) and Amended Protocol II (see Article 4 and Technical Annex). These are examined at pages 47-65 of the paper, where it is argued that they provide a possible basis to require integrated diagnostics, as a corollary to self-destruction and self-deactivation features, and munition detectability for easier clearance in a post-conflict context.
The full paper, Autonomous Weapon Systems, Drone Swarming and the Explosive Remnants of War, can be downloaded on open access from SSRN.