Autonomous Resupply: Drones for the Supply Chain reinforce the Kill Chain

Robots are no longer new on the battlefield, and neither are drones after their widespread use in nearly every major conflict fought by a modern Army in recent times. The ethics, responsibilities and vulnerabilities of military robots are still being debated in academic, industry and government circles whilst robots have added to the capabilities of Intelligence, Surveillance and Reconnaissance (ISR) and, occasionally delivered weapon effects as required. But the most understated potential for drones revolutionising our methods of war is unharnessed yet – the inevitable exploitation of drones to be the backbone of supply distribution in the battlespace.

Compared to the ability of loitering UAVs to reconnoitre terrain, or weaponising an autonomous platform with Hellfire missiles, logistic drones and machines don’t appear to be sexy. There’s not much hype around drones for resupply, in stark contrast to the sensation of drones for the kill chain.

But the kill chain is useless without a resilient supply chain – and this is where, in recent times, the incentive to employ autonomous resupply methods has outweighed the costs for distribution via manned resupply. The modern Army has a Tooth-to-Tail Ratio (T3R) of approximately two logisticians supporting every warfighter – an inverse of the World War I figures where every logistician supported two warfighters. In fact, the T3R is one statistic that has worsened steadily over the last century of warfighting. The improvements in combat technology have not transposed to combat logistics- we still, essentially distribute supplies akin to the horse and cart methods our predecessors had: a human needs to operate a vehicle, and travel with supplies to load and unload the supplies. We are still limited to what the “horse” (vehicles) and human can do as autonomous resupply is yet to reach the military mass market.[1]

While many military projects have toyed with the unmanned, or reduced manning, model of resupply, the pressure is on civilian freight and distributor companies to pioneer the hardware, software and systems to make unmanned resupply relevant to a competitive world of business logistics. Companies such as Amazon[2], Fedex[3], UPS[4] and Rakuten have proven that unmanned distribution is feasible in regulated airspaces. Interestingly, most of  the business logistic theory behind these companies grew from lessons from World War II, coupled with concept of marketing in the 1950s[5]. “However, it does not necessarily follow that business best practice provides the benchmark for military logistics”[6] and this stark warning should both inspire and caution our future employment of drones for military resupply and Humanitarian Aid and Disaster Relief missions.

A quick look at the civilian prospect of logistics drone use is optimistic: PWC says that $13 billion of transport benefit from drones out of the potential $127.3 billion value of the global market, powered by drones in business[7]. Only the industries of agriculture and infrastructure lead the transportation business, although major intersections of all three industries would interest Defence capabilities – especially at the Force Support level where infrastructure is a premium.

There are overwhelming reasons why autonomous resupply, via ground, air or sea vehicles, is worth employing in the modern battlespace. Whether it is to improve the efficiencies of manpower and resources, or minimise the logistic footprint vulnerable to enemy attack, resupply by autonomous distribution should always contribute to the warfighter winning in the battlespace.

The urban battlespace is perhaps where this is most needed, and most relevant. Given our rising urban population in the world[8] and inefficiencies of the traditional distribution method, urban operations will be difficult through this age of warfare. The current doctrine (LWP-G 3-9-6 Operations in Urban Environments) briefly mentions resupply in the Urban Environment[9]:

First-line sustainment is primarily the responsibility of the manoeuvre elements. Resupply by Combat Service Support (CSS) elements is a key vulnerability; movement and entry by CSS will require protection and other support from manoeuvre elements.

Whilst the responsibility of first line sustainment lies with the manoeuvre elements, many would agree that manoeuvre elements are better employed to hold, defend or seize terrain. Moreover, given that we are a small Army, there is a high incentive for more of our manpower to be employed as warfighters rather than in combat support or combat service support roles. This isn’t to say that machines should take over soldiering – rather, the role of an Australian Soldier should be irreplaceable by any machine (think Contract with Australia). Essentially: humans should do what humans do best, and allow machines to do what they do best.

The United States Army is onto this concept already: the Robotics and Autonomous Systems (RAS) strategy[10] outlines a near-term, mid-term and far-term timeline for a range of capabilities that will improve sustainment, first through semi-autonomous means, then eventually with fully autonomous unmanned aircraft by 2040. The United Kingdom is also all ears for ideas: the Autonomous Last Mile Resupply System (ALMRS) competition is having its first collaboration event this month with demonstrations to follow through next year[11]. In the background, some scoping work by Australian Defence Science and Technology members such as Guy Gallasch and Ksenia Invanova have helped contextualise the reality of this timeframe: we should be seeing autonomous, unmanned vehicles reducing the risk to logistic tasks, as we have seen Unmanned Aerial Vehicles reduce risks to ISR missions.

Advocacy for autonomous systems is quite strong in Western armies and the pressure is on as other nations build capabilities for artificial intelligence as well. Elon Musk, the billionaire known for his work in Tesla, SpaceX and SolarCity, says that an artificial intelligence arms race will most likely cause World War III[12]. That certainty may be doubtful amongst other resources that may lead to war, but in the civilian sector, artificial intelligence is projected to put more than 3 million Australian jobs[13] at risk – their skills will be replaceable by machines. These jobs are primarily the labour-intensive, repetitive or high-risk roles – think of workplace health and safety as a vehicle to justify employing a robot over a human, and the training liability that is significantly reduced.

Army is not immune to this movement, as either our adversaries and allies will introduce systems that will be accessible to the civilian market, or we will invest in RAS over coming years. Alongside the introduction of autonomous systems on the battlefield, we will also need to consider the denial of enemy systems that may be similar, swarming or saturating the battlespace to provide redundancy for resupply methods. Although drones have been commonplace over the last decade of warfighting, it is economically not viable to use traditional air defence systems to shoot down drones[14].

For the Army of 2040, the costs will be worthwhile as humans are best employed into warfighting roles, or in support of the array of technology behind the warfighter. Manoeuvre plans will be offered more flexibility with less dependency on the Daily Replenishment Implementation Program (DRIP) – the autonomous resupply model of the future will be resilient and offer continuous contingency resupply plans, dictated through an artificial intelligence, data science model with human oversight. This doesn’t negate logistic planning, but rather forces more analysis upon logistic planners to work with the deep learning neural networks.

We are also forced to rethink traditional methods of conducting tasks such as distribution points and casualty evacuations: Can we trust an unmanned air vehicle to carry a litter patient, stabilised in flight without medical personnel? Can platoon-level call-signs act as mobile distribution points, with combat resupplies flown directly to a cache or the callsign’s non-commissioned officers? Will a swarm of a few dozen drones, each carrying a shoebox-size load, be enough to survive the enemy’s air interdiction en route to the dependency, and what percentage loss are we able to risk?

The challenge is before us to tackle the technological and organisational constraints that will be inherent with this Revolution in Army Logistics within a Joint Logistics framework[15]. Contested electromagnetic environments, adverse weather and cybersecurity will all need to be considered amongst other traditional threats, and we are also continuously learning from civilian autonomous resupply systems such as Zipline’s[16] airdrop missions.

The Army is the right organisation at the right place and time to demand a resilient autonomous resupply systems to augment human warfighters. A small Army like ours demands minimal logistic drag, while ensuring that CSS capabilities are scalable to meet the needs of all the dependencies. The supply chain, if revolutionised, will in turn significantly improve the efficiencies of the kill chain in any level of warfare.

[1] The Kaman K-Max helicopter was successful for the USMC, but this is not scalable: its effective output is still inefficient in comparison to the human effort to maintain the aircraft and its systems.

[2] https://www.amazon.com/Amazon-Prime-Air/b?node=8037720011

[3] https://www.technologyreview.com/s/602896/fedex-bets-on-automation-as-it-prepares-to-fend-off-uber-and-amazon/

[4] https://www.wsj.com/articles/ups-uses-drone-to-deliver-package-to-boston-area-island-1474662123

[5] LWD 4-0, Combat Service Support, page 40, paragraph 7, Annex B to Chapter 1

[6] Ibid

[7] https://www.pwc.pl/pl/pdf/clarity-from-above-pwc.pdf



[8] Population Division of the Department of Economic and Social Affairs of the United Nations, ‘ World Urbanisation Prospects, the 2011 Revision (2012), http://esa.un.org/unup/Maps/maps_urban_2025.htm , p .4

[9] See Future of Resupply in the Urban Environment: A Focus on Megacities by Lieutenant Alex Jack, 2016

[10] http://www.arcic.army.mil/App_Documents/RAS_Strategy.pdf

[11] http://www.arcic.army.mil/App_Documents/RAS_Strategy.pdf

[12] http://www.abc.net.au/news/2017-09-05/elon-musk-says-artificial-intelligence-will-likely-cause-wwiii/8873992

[13] http://www.abc.net.au/news/2017-09-05/elon-musk-says-artificial-intelligence-will-likely-cause-wwiii/8873992

[14] http://www.telegraph.co.uk/technology/2017/03/16/missile-worth-3m-used-shoot-cheap-amazon-drone/

[15] /joint-logistics-by-design-exploiting-the-momentum-of-the-first-principles

[16] https://www.wired.com/story/zipline-drone-delivery-tanzania/