Camel Train and the Australian Way of War: Building a Robotic Logistics Fleet for the North

Wing Commander Keirin Joyce CSC is one of the Royal Australian Air Force’s foremost architects of uncrewed aerial systems capability. Since joining the ADF in 1997, he has served across operations, engineering, and acquisition, including combat deployments to Iraq and Afghanistan, leadership of Army’s tactical and nano UAS programs, and senior engineering roles for platforms including the MQ-4C Triton. A two-time Spitfire Memorial Defence Fellow and Chartered Professional Engineer,

Joyce has become one of Australia’s most prominent advocates for military autonomy and the integration of uncrewed systems into national defence.

In a recent conversation, we explored the Camel Train project, a sovereign Australian effort to build an autonomous logistics fleet across the country’s vast north, and the broader strategic logic behind it.

The Strategic Problem: Dispersion, Deception, and the Australian Archipelago

The conversation began with a problem that has grown increasingly urgent as Australia has sharpened its understanding of the threat environment it faces: how to operationalise the country’s geography as a strategic asset rather than a liability.

The concept of Australia as an “island archipelago” reframes the vast, sparsely populated stretches of the continent’s north not as empty ground but as maneuver space. The large patches of brown dirt are, in Joyce’s words, “metaphorically bits of water in between those cities”, spaces that need not be permanently manned but can be monitored, activated, and exploited through robotic sensors and autonomous systems.

The immediate tactical challenge this creates is one of agile combat employment. Australia possesses only a handful of permanently manned air bases. In a conflict, those bases will be targeted. The solution  —  dispersing combat power across a network of bare bases, civilian airstrips, and even straight stretches of highway  —  is conceptually straightforward but operationally demanding. Moving F-35 packets, forward arming and refueling points, and the full logistics tail of a dispersed air campaign requires a sustainment architecture that is characterised and facilitated by the concept of agile airbase operations.

And there is a second-order challenge layered on top: dispersion cannot be concealed indefinitely. An adversary will be watching for the indicators of movement, the combat signatures of a force element relocating to an agile air base. Joyce’s answer to this is what he calls masking: using robotic systems to generate false signals at decoy locations simultaneously, creating dilemma in adversary targeting calculus. While real F-35 packets move to locations A, B, and C, robotic systems generate signatures that suggest operational activity at locations X, Y, and Z. One way, conceptually, to execute that kind of deception at scale across Australia’s top end, could be facilitated by a large, cheap, and highly mobile fleet of autonomous aircraft. As well as the logistics role that Camel Train is designed to provide, it could do this role also.

The Camel Train Project: Sovereign, Scalable, and Already Flying

Camel Train was announced by Chief of Air Force at the Air and Space Power Conference in 2024 and has moved with unusual speed. Within two years, the program has taken its prototype aircraft, a Jabiru 400 retrofitted with a fully Australian autonomous avionics stack, from conception to flight testing under a CASA experimental certificate.

The architecture is deliberately sovereign and deliberately modular. The core avionics stack combines a civilian flight control system from CubePilot in Geelong, an optical detect-and-avoid system from Revolution Aerospace in Brisbane, and a radar-based detect-and-avoid system under development from Mission Systems in Sydney. These components have been integrated under the engineering leadership of RF Design, also in Brisbane. The result is a system that can move approximately 200 to 250 kilograms of logistics over planning distances of 750 kilometres, useful military loads including jerry cans, ration boxes, and ammunition tins.

But the more important design decision is that the avionics stack is aircraft-agnostic. The same package that has been installed in the Jabiru 400 can be pulled out and installed in a Cessna 152, a King Air, or a Gipps Aero GA8. The logic is straightforward: standardise the avionics, then let the airframe selection be driven by mission requirements and availability.

This leads directly to what is perhaps the most striking element of the program’s strategic concept. There are approximately 50,000 general aviation aircraft in Australia. If the Camel Train avionics stack can be produced as a retrofit kit and installed in civilian aircraft, then mobilisation in a competition or conflict phase does not require building new airframes. It requires acquiring aircraft that already exist and installing the kit. As Joyce observed, this is not entirely unlike what Ukraine did in 2022, taking high-wing general aviation aircraft, installing control systems, and deploying them for long-range missions. The methodology is analogous, though the Australian application is oriented toward logistics sustainment rather than strike.

Fleet Architecture: From Platform Thinking to Payload Thinking

One of the conceptual contributions that emerges from Camel Train is a fundamental reorientation away from platform-centric procurement toward fleet and payload thinking.

The traditional defence acquisition logic asks: what platform do I buy, and how many? Camel Train inverts this. The question is: what outcome do I need, and how do I build a fleet of interoperable, scalable systems to deliver it? The avionics stack is the common thread. The airframes are variables. The payloads, logistics loads, ISR sensors, SIGINT packages, decoy emitters, can be mixed and matched across the fleet depending on mission requirements and basing context.

This has significant implications for how the fleet is managed. Joyce described a Camel Train management system and accompanying ground control station that already simulates twenty Jabiru 400s operating concurrently, assigning logistics tasks across the country. Scaling that simulation up is a software problem, not a hardware one. The vision is a one-to-many control architecture in which a ground handler loads cargo into an aircraft, presses a button, and the management system handles routing, sequencing, and coordination across dozens or hundreds of aircraft simultaneously.

At the smaller end of the spectrum, Camel Train has been prototyping alongside complementary systems: the Sypaq precision payload delivery system, the so-called cardboard drone, and the ALADDIN, a 30-kilogram multi-rotor designed to be deployed from air mobility aircraft, with rotors that extend autonomously upon release and the aircraft flying itself to a destination. Both run on CubePilot autopilots, maintaining the common control architecture that the fleet concept requires.

ISR, the Kill Web, and the Commercial Technology Ecosystem

A logistics drone flying across Australia’s north is not just a logistics drone. Once you accept the modular payload logic, every mission becomes a potential intelligence-gathering mission as well. Joyce was direct about this: “a boring old logistics drone can turn into a high end contributor to the kill web.”

Australia has developed a genuinely impressive cluster of small companies in lightweight, low-altitude sensing, Sentient Vision Technology (now part of Shield AI), Arkeus Systems, AVT Australia, C2 Robotics, SYPAQ, and others. Their electro-optic systems are already commercially successful and globally exported. Adding a few kilograms of optics to a Camel Train aircraft turns each logistics sortie into a passive ISR pass over the archipelagic region. Aggregated across hundreds of sorties and pushed back over communications networks to intelligence centres on the Australian mainland, that constitutes a persistent, low-cost surveillance capability that would be expensive and manpower-intensive to replicate through conventional means..

This points to a broader observation about the nature of defence-relevant technology in the current period. The assumption that Defence must be at the leading edge of ISR technology has eroded. The commercial sector driven by civilian aviation, agriculture, logistics, and geospatial industries — has produced capable, proven, affordable systems. The defence task is increasingly one of integration and orientation: identifying which commercial capabilities can be bent toward military purposes, and building the programmatic structures that enable that leverage.

A Whole-of-Society Defence Structure

Perhaps the most ambitious aspect of the Camel Train concept is not technical but social. Joyce pointed to a figure that frames the workforce question in striking terms: there are approximately 30,000 licensed drone operators in Australia, all certified under CASA standards. These are not military personnel. They are commercial operators, hobbyists, agricultural technicians, and surveying professionals. In a competition or mobilisation phase, this workforce represents a potential operational reserve that has never existed before.

The logic extends further. Many of the missions that a Camel Train fleet would execute, routine logistics runs, passive ISR passes, decoy deployments, do not require the training and expertise of a uniformed military pilot or UAV operator. Those skilled personnel can be reserved for high-end combat missions. Logistics missions can be handled by a civilian workforce, operating from the relative safety of southern Australia and flying missions in the north through secure communications networks.

Joyce framed this as a “fifth generation workforce” concept, a deliberate play on the language of fifth generation air combat, applied to the human domain. The geography of Australia, which in some respects creates the problem of dispersion, here becomes an asset: operators in Adelaide or Canberra are well outside the threat envelope, while their aircraft operate across the top end.

This suggests a defence structure that is genuinely whole-of-society rather than whole-of-government in the conventional sense. It is not simply about government agencies coordinating. It is about tapping the commercial sector’s capabilities, civilian aviation’s workforce, and the country’s distributed engineering talent in ways that make the distinction between civil and military capability more fluid and that give the broader population a concrete stake in and contribution to national defence.

Leveraging What Exists

The final theme of our conversation was perhaps the most important strategically: the difference between building new capability and leveraging existing capability.

Camel Train did not emerge from nothing. It has been made possible by fifteen years of NATO standards work on uncrewed aircraft systems and data interchange standards for electro-optics, infrared, and synthetic aperture radar, work that means sensors from different manufacturers can exchange data packets across services and nations. It has been made possible by a decade of Australian government investment through the Defence Innovation Hub in optical sensing companies that are now commercially mature and globally competitive. CubePilot has sold millions of line items worldwide. RF Design has already exported six figures worth of radio modems. The technology ecosystem was not built for Camel Train but Camel Train is the programmatic structure that harvests that ecosystem’s output.

Joyce made an observation that resonates well beyond this program: Australia is never going to operate at the production scale of major powers. But there are specific things Australia does exceptionally well, robotics and autonomous systems, precision sensing, advanced engineering in niche domains. Those are also the things that happen to be most relevant to the emerging character of warfare. Investing further in those strengths, rather than attempting to replicate the full-spectrum capabilities of much larger allies, is both strategically sound and financially realistic. Australia could be producing hundreds of robotic uncrewed systems per year from its existing small but highly capable design and production houses. That is a meaningful contribution to coalition capability, not a second-order one.

The mental shift required is from thinking about the military-industrial complex as the source of defence capability to thinking about leveraging strategies that connect military requirements to the commercial and civilian sector’s natural outputs. Civil aviation exists in Australia because the country is large and sparsely populated, not because of defence planning. But that fifty-thousand-aircraft civil aviation sector is now a potential strategic asset, waiting for the institutional imagination to make use of it.

Camel Train is early evidence that imagination of this kind is available. Its significance lies not only in what it will deliver for agile combat employment in Australia’s north, but in what it demonstrates about how defence capability can be built: modularly, collaboratively, at speed, and in ways that the whole of Australian society can contribute to and take ownership of.

For Joyce’s introduction to my recent book on the impact of drone dynamics and development on defence, see the following: