Proteus: The story so far

A little over three years into a four-year contract for the Rotary Wing Uncrewed Air Systems (RWUAS) Phase 3A Technology Demonstrator Programme, referred to as ‘Proteus’, we have made significant strides in the design and development of the circa three-tonne uncrewed rotorcraft which will demonstrate advances in autonomy and payload modularity and interchangeability, and the viability of Large Autonomous Vertical Take-Off and Landing Uncrewed Aircraft Systems (VTOL UAS) on military maritime operations.

Why autonomy?

As the Royal Navy looks to deliver its Maritime Aviation Transformation (MATx) strategy – which covers the evolution of maritime aviation until 2040 – it is doing so against the backdrop of shrinking budgets and a reduction in force, as well as growing concerns relating to the threats posed by hostile nations.

The growing strategic importance of autonomy is reinforced in the recently published Strategic Defence Review which stated that “drones, AI and autonomy” would complement aircraft, ships and artillery as part of the “new vision for how our Armed Forces should be conceived – a combination of conventional and digital warfighters.”

The significant challenge of recruiting and retaining military personnel can also be addressed by introducing greater autonomy to the Royal Navy’s portfolio of technology and activities.

Autonomous platforms bring potential cost benefits vs crewed platforms. For the former, you get more capability as they require less mechanical complexity and avionic equipment. Furthermore, there is no crew controlling the aircraft, so mechanical control runs, seats and avionic displays are all unnecessary.

Additionally, because the aircraft has no crew, the design approach can be slightly less orthodox; the crashworthiness of your design, for example, will consider different factors when there are no longer humans on-board- the aircraft.

Autonomous aircraft are viewed as a key enabler for the future of maritime operations, with crewed and uncrewed assets flying and operating alongside each other.

Proteus will enable the Royal Navy to deploy mass for long durations for dull, dirty and dangerous missions. Searching for submarines in the dark, in the rain, in heavy North Atlantic seas for hours and hours at a time is a challenging place to be. If an uncrewed aircraft can do that itself – by staying focused on the task and having longer endurance since it’s carrying more fuel rather than people – that will be a significant and positive advancement for how the Royal Navy conducts such exercises.

Seeing is believing

September 2023 represented a real milestone for us, as we pivoted from autonomous flight software development developed from existing Leonardo products to the new Proteus mission system platform requirements.

During the following ten months, our cross-domain team was formed and it accelerated building autonomous mission capability through a spiral development process.

This allowed us, in July 2024, to deliver the first synthetic demonstration of an Anti-Submarine Warfare (ASW) FIND mission, including locating and classifying the submarine using a single aircraft.

It provided the MOD customer with a tangible visual experience of the aircraft in a synthetic environment and how it would operate using task-based autonomy.

Layering complexity

Most recently, in May 2025, we undertook a further demonstration introducing objective-based automatic mission planning, with task management and the handover of tasks between three aircraft in the synthetic environment. This featured wide area surveillance, consolidating a multitude of sensors including electro optical, radar and AIS data.

Through data fusion, the aircraft established the situational awareness picture of what was around it, ultimately allowing it to feedback to a ship combat management system and giving that tactical picture back to the Royal Navy command to inform their strategic decisions.

Human v machine

The latest demo saw the aircraft behave in ways that are new to us, because it is making decisions autonomously and that are jarring to a human to observe. Examples include the on-board decision to hand over a task it cannot achieve to another aircraft without operator input which that aircraft accepts and undertakes, non-repeating cover breaks (flying off in different undefined directions) or automatic route re-tasking based on sensor feedback.

We fully expect to see more of this learned behaviour from the aircraft; it’s different to how a human would operate, and that’s good. We need to embrace that. Just because we’ve always done it that way, doesn’t necessarily make it the right way to do it.

It also means that the aircraft can manoeuvre in ways that might not be physically comfortable for a human being – like rapidly turning and changing course frequently. For a pilot, it’s an unpleasant way to operate, but the aircraft doesn’t care about that; it’ll just do whatever is the most efficient. Furthermore, for a pilot there’s human nature at play. They’ve been trained, so it becomes built-in behaviour; they always do it that way, though it doesn't necessarily make it correct.

Moving forward, we are embracing the results we are seeing in these exercises, trying different ways of operating in our synthetic environment, which is a great facility to validate this. Undoubtedly, there is huge excitement in the team as we continue to learn more about what autonomy can offer – not only in delivering the mission objectives we had expected, but how the aircraft's independent decision making can deliver more effective results.