Army Rumour Service

Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!

£30m injection for UK’s first uncrewed fighter aircraft

The UK’s first fleet of uncrewed fighter aircraft is one step closer to reality following a £30 million contract to design and manufacture a prototype in a three-year deal supporting more than 100 jobs in Belfast.

Continue reading...
 
The UK’s first fleet of uncrewed fighter aircraft is one step closer to reality following a £30 million contract to design and manufacture a prototype in a three-year deal supporting more than 100 jobs in Belfast.

Continue reading...

‘....these aircraft will be the UK’s first uncrewed platforms able to target and shoot down enemy aircraft...’


Unpossible! ARSSE’s finest minds told me so!
 
Interesting how little of the article is about the capability, and how much is about generating jobs in NI.
 
Britain’s Loyal Wingman will be a seriously advanced project, building the drone itself not quite so problematical as the ‘brains’ or AI that goes along with it. August 2020, US’s DARPA AlphaDogfight trials in a series of AI experiments in simulator fighter cockpits an Air Force F-16 pilot lost five dogfights in a row against an AI pilot in virtual combat.

DARPA’s Air Combat Evaluation (ACE) program is working on the digital assistant or ‘ brains’ for future drone/loyal wingman/ autopilots. This is an ongoing and extremely complex process. Unlike humans AI doesn't have an instructor, a technical manual, or previous experience on how to fly. It needs huge amounts of data on how any particular aircraft actually flies and particular type's aerodynamic data on speeds, G-loads, angle-of-attack and stall/spin characteristics, and a host of other aero/control data to learn that skill from.

The loyal wingman concept is advancing along with the AI that governs the drone’s semi autonomous flight abilities and reactions to enemy actions and reactions during combat, as well as advancing the abilities of present autopilots. These learn to fly and dogfight like a human does, first by exploring how to control a particular aircraft and then how to manoeuvre against another simulated flying machines. Engineers are using a training architecture called proximal policy optimisation which is a method of machine learning in AI for obtaining optimal solutions for the tasks involved.

The objective with the AI side is developing a system that can fly an unmanned aircraft with minimal, if any human interaction. Possibly using a software that is “modular” and “open architecture” so that new “complex autonomous behaviours” can be rapidly added into this advanced computer brain in the future.

Spirit AeroSystems has been selected to develop and manufacture a prototype of the UK's loyal wingman drone, and awarded a three-year, £30 million contract to start the development at its branch located in Belfast. Project mosquito also involves Northrop Grumman’s UK division.

 
Last edited:
‘....these aircraft will be the UK’s first uncrewed platforms able to target and shoot down enemy aircraft...’


Unpossible! ARSSE’s finest minds told me so!
Two points, but I'm not sure how much you comprehend them:
  • that line comes from the PR person who wrote the press release, and is not a direct quote from the staff or engineers involved
  • "uncrewed" is very definitely not the same as "autonomous". Having all the radar emissions coming from a UAV a few miles away*, and the resulting data handed back over a datalink; is an incremental step in survivability, not a huge leap forward in warfare.
* It puts a price on aircrew lives, mind you - yes, you can use a smaller / cheaper radar if it's going to be closer to the enemy; but a decent fighter aircraft radar isn't a cheap piece of kit (twenty years ago it was seven figures, unless you're mass-producing them), and very definitely isn't something you regard as an expendable item.
 
From Ace Jewell, Commander USN (retired), now 88 years old, fighter pilot in 3 wars:

"Drones will not be late to briefings, start fights at happy hour, destroy clubs, attempt to seduce other pilots' dates, purchase huge watches, insult other services, sing ‘O'Leary's Balls,’ dance on tables, yell ‘Show us yer tits!’ or do all of the other things that we know win wars.

I see no future in them.”
 
From Ace Jewell, Commander USN (retired), now 88 years old, fighter pilot in 3 wars:

"Drones will not be late to briefings, start fights at happy hour, destroy clubs, attempt to seduce other pilots' dates, purchase huge watches, insult other services, sing ‘O'Leary's Balls,’ dance on tables, yell ‘Show us yer tits!’ or do all of the other things that we know win wars.

I see no future in them.”
Or any of those other reflex, instinctive things that humans are capable of that a computer would never think to try that can tip the balance the right way
 
But on a more serious note the use of ‘Loyal wingmen’ is a serious move forward in Electronic Warfare. Until the last century traditional combat methods were the norm. Attacks and defence, soldier to soldier, ship to ship, plane to plane and the various combinations. Electronic warfare had made an appearance in at least the beginning of the 20th century. The earliest documented consideration of EW during the Russo-Japanese War of 1904–1905 and grew from then on. As technology became more available and affordable, adversaries grew their EW capabilities. Now, these are growing more sophisticated and... unpredictable, due to their ability to "learn" and adapt.

With AI, intelligent machines work and respond a lot like humans, but machines are performing smarter tasks, faster. Machine learning takes AI one step further, allowing machines to continuously learn from data and adapt as a result. Computers are learning at a very rapid rate. Threats using machine learning learn from every encounter, determining ways to be more effective so that they can overcome countermeasures.

This is happening without human interaction, as the computer decides how to alter behaviours When tested or engaged, these threat systems learn from that experience then modify their future behaviour, ie the computer is deciding the next steps. Due to the system’s unpredictable behavior, even people responsible for the system cannot predict its exact behaviour.

As threat systems advance with machine learning technology, they adapt and alter their behaviour/course of action at an increasingly rapid rate. An example might be a radar trying to track a jet. The adversary’s countermeasures may stop it from succeeding but using machine learning the radar would repeatedly try new approaches to overcome these. Today’s machines possess intelligence that is an order of magnitude higher than a human expert in EW and as they learn that data continues to aggregate.

Responsive threats already exist, often called cognitive and adaptive. Although these terms are used interchangeably, levels of adaptability exist. Most of them do not come near the capabilities of cognitive EW. Using machine learning, cognitive EW systems can enter an environment with no knowledge of an adversary’s capabilities then by doing something that makes the adversary’s system react, they evaluate its response, quickly understand the scenario, then develop an effective response suited for that particular adversary’s system.

In contrast, adaptive solutions cannot rapidly grasp and respond to a new scenario in an original manner. An adaptive radar can sense the environment then alter transmission characteristics, providing a new waveform for each transmission or adjusting pulse processing. This flexibility allowing it to enhance its target resolution, for example. Many adversary systems require only a simple software change to alter waveforms, which adds to the unpredictability of waveform appearance and behaviour leaving adversaries trying to isolate adaptive radar pulses from other signals and whether it’s friend or foe

But the EW domain is just beginning to implement machine learning and AI. These technologies and their applications will evolve rapidly resulting in increasingly complex threats and 'loyal wingmen’ with AI will become increasing needed on the battleground, air land or sea.

 
Top