Sustaining Machines: Logistics and Autonomous Systems

By David Beaumont

The popular discussion on autonomy in warfare is constrained to either describing the advantages of introducing autonomous systems for ‘dull, dirty and dangerous’ work, or articulating the limitations of their use (including ethical limitations).

In terms of ‘logistics’, we can focus on how automation promises to distribute more things to more combatants more quickly, replaces forces in the field, or help us to be more productive and economical with our resources.

Automation offers military logisticians tremendous advantages and has to be a part of their future.

The opportunities for automation in logistics are virtually limitless, only requiring technology and entrepreneurship to deliver a generational change to military technology.  It becoming easier to find such opportunities given the vibrancy of the industry sector, and the enthusiasm robotics presently generates in defence circles. Rather than go through all of these opportunities, this article describes the capability that brings all of the pieces of an autonomous logistics system together – what we call the control network.

This is a strategic capability which must be invested in.

Then I will discuss: What does the introduction of autonomous systems mean with respect to generating and deploying military forces?

This is an important topic because to properly introduce autonomous systems into the military, we best be prepared for organisational change, cultural change and necessarily closer relationships between the military and industry.

In other words, I’m going to talk briefly on how we might make the capabilities we intend to develop practically useful and sustainable.

Automated and autonomy in logistics

The military use of autonomous systems conjures the vision of multi-domain technologies connected together in a mutually-supporting ‘kill web’. Swarmed drones, larger UAS, submersibles and other capabilities operating automatically and nominally without human influence (maybe even interference).

Though the technology is revolutionary, the idea is not; the “kill web” is to combat operations what the logistics control network is to framework which sustains the operations.

This network uses sensors to make decisions about what is moved where.

As militaries introduce more and more autonomous systems into service, many of them to fulfill logistics tasks, the importance of this network cannot be understated.

We’ve had an automated logistics control network for decades.

Logistics information systems – with all their benefits, liabilities and risks – have been essential to commercial and military logistics since the invention of computers. They have allowed the archetypal complex system – the commercial supply chain – to be analysed to excruciating detail. There is nothing stopping militaries becoming an ‘Amazon’ given the technology that is on offer; provided it is militarised and reflects the needs of fully deployable system that can function in a crisis.

Naturally, as armed forces explore the use of autonomous systems, they will also have to explore the use of automation to truly leverage the opportunities autonomy will deliver.

Let’s militarise the idea.

Effective logistics information systems enable these forces to more efficiently prioritise and allocate resources by leveraging sensor-based analytics, thereby creating additional capacity in the military supply chain and other logistics functions.

When greater logistics capacity is found, this naturally means more options open up for the strategist, tactician or capability manager. Incorporate scalable, swarming, logistics UAS or autonomous convoys into this system and a remarkable level of efficiency might be possible.

The use of information-age technology has helped address what has been described as ‘the logistics snowball’ – the propensity of poorly planned and executed logistics to expand logistics requirements as more and more people, and more and more resources, are directed to problem solving.

The opportunities on offer to us with future forms of autonomous systems are tremendous and will undoubtedly continue to be exploited.

But we must ensure that whatever logistics autonomous systems are introduced that they are introduced with a backbone control network that makes the whole effort worthwhile.

Automated, not autonomous, logistics is probably where the best return of investment of the defence dollar lies.

Automated systems remove the guesswork out of logistics – militaries can get a truer sense of our capability and capacity at any given point in time. Vehicle ‘health and usage monitoring systems’ and other similar technologies are exemplars of this. They enable decisions about capabilities to be made at a faster tempo than ever.

It’s been a rocky journey with the systems – for example, the ‘Autonomic Logistics Information System’ for F-35 Joint Strike Fighter has received a significant upgrade to overcome highly-publicised problems – but this really is a new era of information management and problems are inevitable.

For these systems to offer the most to military logisticians, there is the issue of data management that we must eventually come to terms with – who owns it, when it can be used and for what reason – including ownership of the algorithms that may ultimately make decisions which were formerly the purview of military commanders.

Information, extrapolated into data, provided by these systems, is strategically vital. The complexity of military supply chains has expanded with globalisation, increased civilianisation and outsourcing of logistics capability, and with the sharing of capability across coalition partners.

Automated logistics, appropriately secure, will help us garner where risks lie such that timely plans can be developed. Shortages could be better avoided. Costs could be better understood. Supply through multiple levels of producers and manufacturers can be accurately tracked thereby alerting the military to risks relating to the manufacture of capability.

Secondly, autonomous systems may have the computational power to predict and automatically react to ensure the right product is at the right place at the right time. This will assist in signalling industry as to where supply deficiencies lie, and can support mobilisation processes when strategic crises first appear.

It is important to be aware of the risks.

Cyber threats are persistently targeting global businesses, so Defence must prepare itself during the transformation of its logistics capability. In a 2018 testimony to the US Senate, the Commander of US Transportation Command General Darren McDew, highlighted the cyber threat to logistics as ‘being the greatest threat to our military advantage.’ Malicious state and non-state actors are already targeting vulnerable, largely unprotected, commercial systems linked in with barely protected military logistics systems. This threat was verified in the Defence Science Board 2019 report on ‘Survivable Logistics’.

Why would a hostile target a hardened, highly classified decision-support and command and control network, when a soft underbelly is already presented to them?

Though there are threats to the automation of our military supply chains, the positives clearly outweigh the negatives. It is unequivocally the best solution to the logistics problem of our time – productivity. Logistics autonomy fundamentally gives armed forces greater capacity to do more with less, or better still, much more with the same.  It simplifies something that would otherwise be highly manpower intensive.

But think of what can be done when an advanced predictive AI is aligned with a scalable autonomic distribution system?

It can provide new vectors to deliver battlefield resources to the point at need, at a lower risk to human life. There are considerable financial advantages to Defence and Government if such capabilities are programmed and funded, and military advantages that might just contribute to the elimination of the large logistics footprint within an operational area.

The Logistics of Autonomy

How are forces that include autonomous systems generated and sustained?

Militaries using autonomous weapons will, if we are optimistic about the technology, necessarily look very different in twenty, thirty years in the future. It is largely self-evident that bringing new technology into military organisations is challenging.

What isn’t often acknowledged is how impactful this technology may be on the characteristics of the military organisation. The introduction of technology can have hidden consequences which are rarely apparent until the technology is in use.

New ways of doing business will be needed, organisations redesigned and policies created. Naturally, it’s important that these outcomes are prepared for.

Firstly, military logistics in war will be different – it is going to change in a way that hasn’t been seen since the combustion engine was introduced. Motorisation, mechanisation, rocketry and flight have already elevated the importance of specialist mechanics, petroleum operations, munitions specialists and supply specialists. So too has the act of providing better materiel and training to each combatant – the battlefield has been ‘thinned’ with each able to bring greater firepower to bear on the enemy than the previous generation, but the logistics per unit cost of the combatant has also increased. Conventional military forces are not getting logistically ‘lighter’, and despite the desires of hopeful force designers, are unlikely to do so with automation.

The centre of gravity for military forces is inextricably moving from the battlefield and to the supply depots, bases, ports and defence infrastructure – the ‘rear echelons’ – as new technologies such as autonomous systems beckon.

We’ve got to appreciate what this means in the context of opportunities, risks and vulnerabilities.

Secondly, the introduction of complex systems and machines will transform the way militaries will organise.

I wrote here, citing Chris Demchak, of the introduction of the M1A1 Abrams tank to the US Army forty years ago as an illustration of the problem.

It’s good analogy on the impact of technology on military logistics systems and organisations writ large. This was a tank designed to leverage technology and be less personnel-intensive, with a lower maintenance bill, simple to operate and decisive in combat. It was simple to operate, but proved difficult to repair and sustain. It could not be repaired without the OEM involved, and the technology often mystified even them. The US Army responded by procuring new testing equipment, and created new technical specialities to handle repair.

The level of technology required for the tank made the supply of parts for it challenging. Unexpected costs and the insufficiency of repair parts to support the tank ensured supply was handled with such scrutiny that an entire logistics bureaucracy was created. This generational change in equipment meant that US Army now has an incredibly effective tank with no real peer, but it was not an easy introduction into service.

Militaries will have to prepare for the same with the introduction of autonomous systems.

This will not be a venture without significant implications for the transformation of the militaries over coming decades.

It will not be a venture that can be rushed into without understanding the risks, and we must recognise that we’re at the beginning of this challenge. This challenge will require Defence to experiment, discuss about and tinker with these truly revolutionary capabilities, but it must also consider new concepts and policies to better integrate these capabilities into the organisation.

Thirdly, the logistics liability for operating these systems must be understood.

Separating military robots and battlefield automation from the rest of the discussion, it’s pretty clear that we’re at the infancy of tactically useful systems that can be employed en masse.

At the moment, and especially in the Land Domain, many battlefield systems are ‘brittle’, not particularly adaptable and easily break down. This reflects the difficulty for machines that lack the maneuverability of a human being, or have difficulties operating in close proximity to them.

The situation is better for military aviation and in the maritime context. It will be some time yet before the ‘medic’ is replaced with the ‘mechanic’. But when they are, militaries will have to be respectful that the act will be transformational in the military workforce – especially with Army where this problem will be acute. It may even be aggravated by a human-machine ‘teaming’ approach where both forms of combatant are employed.

Perhaps we can combat the ‘less-positive’ effects of automation by focussing on the notion of disposable military robots.

It’s tempting to think that robots can be abandoned when it is damaged or no longer in use; it appeals to our sense that there is a real possibility that humans can be removed from danger and replaced with something of lesser value. It’s patently a present day unreality save in very small-payload logistics operations. Until production lines run so large that costs are driven down, or newer technologies are found that dramatically lower costs, it will be inevitable that we treat autonomous systems with the same level of care we do any other form of exquisite technology. Nonetheless, it is likely that this problem will be overcome with time, technology and effort.

Fourthly, the link between the military and industry partners will necessarily be closer than ever.

Larger logistics requirements do not always require larger military logistics organisations, but it does mean militaries need to be better at incorporating civilian resources into their operations. Military logistics always extends into the economy – more specifically the nation’s industrial base – and autonomy necessarily means that the integration of industry into the routine sustainment of new capabilities will remain important. It is quite clear that industry partners will have to continue to work closely, if not intimately, with armed forces to provide the technical support and expertise that is traditionally difficult for the military to generate independently. The workforce both generate is one that is shared irrespective of whether a uniform is worn or not. It is also clear that a conversation about how skills may be transferred into the military workforce if needed in a crisis must be had, or how autonomous systems might be sustained and repaired in conflict zone.

This leads onto the final topic I wanted to cover on the logistics of autonomy.

It is not lost on most readers that there is a tremendous opportunity for defence industry to step into an electronics industry gap is only beginning to be filled. If we are to embrace the use of autonomy in militaries such as the Australian Defence Force as a credible alternative to the human combatant, it will be highly advantageous for the military to have a national industry behind it. A dependency on foreign components and construction can become a strategic risk – especially as global supply chains are contested or limited resources shared.

I suspect that electronics and componentry join ammunition and fuel as a marker of strategic resilience in due course. In the meantime, all will need to be careful about accelerating into autonomy else we embark upon a costly sham with unviable capabilities in combat.

Perhaps this necessitates us having a conversation about innovations and their identification as a matter of strategic value and a target of regulation.

Most innovations in autonomic systems will come from the private sector, and in many cases, will be available to the highest bidder.

This could even discount any investments that Governments may make into the sector. A pessimistic view of the future suggests we need preserve whatever advantage we can, and – as a nation – we might have to balance our commercial and strategic interests. Australia’s stake in autonomous systems development is an important one, with strategic implications. With autonomy firmly on the horizon for the ADF and other advanced militaries, it seems clear that we must initiate this discussion now.

This article, I admit, is a smorgasbord of ideas.

Logistics is first and foremost about practicality, and it is important to ensure autonomous systems that are introduced do what is intended.

There’s good reason for an investment in autonomous systems in the short term; they certainly offer a way to overcome some of the structural shortcomings faced by the military in terms of ‘mass’ and an ability to operate as efficiently and effectively as we need to be.

But we must not race to failure.

Militaries cannot afford to let autonomous systems become a capability ‘drag’ by not being diligent.

Automation will create new options at all levels of war, improve the capacity of a defence force pressured by its relative size, and give us new opportunities to exploit. The technology behind automation is an area where Australia can generate a strategic advantage if it chooses to; we have a high standard of education and a long track-record of innovation as a nation.

Western militaries, in general, prepared for change and actively seeking partners to overcome many of the challenges, and take advantage of new opportunities, that have been raised here.

We all know how rapidly the technology around automation is evolving.

The all work on overcoming the logistics limitations of autonomy the better. Technical and conceptual discovery must occur at pace. This way the potential of the technology will be realised, rather capabilities or systems that are too exquisite to be practically employable, unsustainable, or offer little in being part of a strategic offset, result.

This is an enduring problem with introducing new technology into defence forces in a time of relative peace, where there is always a temptation to make expedient decision and mortgage the future as a consequence.

This was adapted from Col. Beaumont’s presentation at the Williams Foundation Seminar on Next Generation Autonomous Systems on April 8, 2021.

The article was recently published on the website Logistics in War.