Behind the scenes, the UK is developing a range of military autonomous systems. Image: Crown Copyright Behind the scenes, the UK is developing a range of military autonomous systems. Image: Crown Copyright

UK military researching armed autonomous drone swarms

 XLUUVs, Swarms, and STARTLE: New developments in the UK’s military autonomous systems

In November 2018 Drone Wars UK published ‘Off The Leash’, an in-depth research report outlining how the Ministry of Defence (MoD) was actively supporting research into technology to support the development of armed autonomous drones despite the government’s public claims that it “does not possess fully autonomous weapons and has no intention of developing them”.  This article provides an update on developments which have taken place in this field since our report was published, looking both at specific technology projects as well as developments on the UK’s policy position on Lethal Autonomous Weapons Systems (LAWS).

The development of autonomous technology remains a research priority for the MoD and the UK’s armed forces.   Research work continues in all three of the key disciplines underpinning autonomous technology – artificial intelligence and machine learning; robotics; and sensors.  Over the eighteen months since ‘Off The Leash’ was published MoD programmes for pioneering new autonomous military systems have increased not only in the breadth of their scope but also in their profile within MoD’s communications.

Click to open report

It is important to note that none of these projects are intended to directly develop a lethal autonomous weapon system – a ‘killer robot’ able to select and destroy targets without any human intervention.  However, they contribute towards an increase in the UK military’s overall warfighting capability, and more strikingly, they represent developments in technology which in due course could be combined with other systems to form the building blocks of a lethal autonomous weapons system.

MoD’s research and development work on autonomous systems covers all three traditional battle domains – air, land, and sea – as well as cross-domain programmes and a broader range of projects to explore military applications of artificial intelligence and sensor technology.  As well as conventional style research projects, MoD’s programmes for exploiting autonomous systems have also included in-the-field evaluation exercises and autonomy innovation accelerator schemes for the Royal Navy and the Army. As well as the developments outlined below, there are almost certainly research projects which are being advanced secretly, behind closed doors.

In the air

Looking first to the air domain and drones, much of MoD’s recent development work has focused on delivering and evaluating a swarming drone capability.  The Defence and Security Accelerator (DASA) scheme, which aims to find and fund new technologies to provide the UK with the most innovative defence and security capabilities in the world, has over the past couple of years been running a ‘Many drones make light work’ research programme to investigate the military uses and practicalities of operating a swarm of 10 – 20 low cost drones which could work collaboratively.  A contract has now been awarded to a consortium led by Blue Bear Systems Research Ltd to deliver live flight drone swarm demonstrations to the military.  As is often the case with such projects, the consortium consists of a mixture of established military contractors, smaller technology research companies, and academia: in this case IQHQ, Plextek, Airbus and the University of Durham.

In February 2019 the then Defence Secretary Gavin Williamson announced plans to use the Defence Transformation Fund to develop “squadrons of network enabled drones capable of confusing and overwhelming enemy air defences.”  Over the following weeks further details emerged, and it was revealed that the Ministry of Defence intended to launch a three-year programme to procure the new drones and reform the RAF’s 216 Squadron as an experimental unit dedicated to developing drone swarm technology.  Although a deadline to establish the new squadron by the end of 2019 was missed, it is now up and running and its work will assist the RAF in working out its future drone swarm requirements and act as a stepping stone towards the widespread use of autonomous drones.  Through 216 Squadron the RAF will be able to develop operating concepts and tactics for swarming and autonomous drones before establishing a full operating capability.

Military commentators have speculated that the proposal to establish a new squadron to operate the drones suggests that they would be larger than the experimental swarming systems that MoD has already been developing, and are likely to fly as swarms of “loyal wingmen” accompanying crewed aircraft to provide protection and surveillance capabilities.  Since 2017, the RAF’s Rapid Capabilities Office has been exploring options for a low cost, low maintenance drone as part of the Lightweight Affordable Novel Combat Aircraft (LANCA) programme, and LANCA may form part of the new squadron’s equipment – possibly operating in partnership with fast jets such as the Typhoon, F35 Lightning, or the next-generation Tempest aircraft which MoD is hoping to develop.  Contracts for phase 1 of the LANCA programme were awarded last year to three teams led by Blue Bear Systems Research Ltd, Boeing Defence UK Ltd, and Callen-Lenz (Team BLACKDAWN partnered with Bombardier Belfast and Northrop Grumman UK Ltd) to produce a preliminary design for the drone, with the hope that a demonstrator test aircraft would be ready to fly by 2022.

Another high-profile area of research relating to drones and unmanned systems in the aerial domain is counter-drone technology, intended to address threats from small drones.  Through a DASA funding competition, MoD has been looking to develop new defence systems which draw upon autonomous decision-making mechanisms and networked sensors capable of detecting, tracking, and destroying hostile drones.  In December 2019 DASA awarded £2 million to a total of 18 projects for developing the necessary technology, including using machine learning and artificial intelligence to automatically identify drones using information from sensors, and methods for stopping drones using interceptors and electronic techniques.

At sea

The Royal Navy is also playing an active role in the development of military autonomous systems.  Having hosted the Unmanned Warrior exercise in Scotland in 2016 to demonstrate the use of maritime unmanned systems, the Navy is now working through the DASA programme to develop unmanned submarines and ‘intelligent ships’ and has set up NavyX, its own autonomy innovation innovator intended to drive forward the use of autonomous systems in all maritime environments.

DASA has organised a funding competition to help the Royal Navy test options and develop expertise in operating extra-large unmanned underwater vehicles (XLUUVs).  At the beginning of March 2020, a £1 million contract was awarded to MSubs Ltd to build a test submarine for exploring the potential capabilities of large uncrewed underwater vehicles in the future.  MSubs will be refitting an existing commercial crewed submarine with sensors, detection capabilities, and autonomous control systems for use as a test platform to understand the capabilities of an autonomous submarine.  The Navy hopes in future to use such submarines, which could remain underwater for up to three months, for covert intelligence gathering and to identify hostile targets and report findings back to a base station.

Following a previous DASA competition, the Navy is currently trialling Roke Manor Research’s ‘STARTLE’ situational awareness software which continuously monitors and evaluates potential threats using a combination of artificial intelligence techniques.  DASA has now broadened the scope of its work to fund an ‘intelligent ship’ competition for proposals to extend the use of intelligent and automated technologies in the warships of the future.  At the beginning of this year nine demonstrator projects across the automation, AI, and human-machine teaming sectors were awarded a total of £1 million with the aim of using AI techniques to overcome ‘information overload’ faced by crews who use increasingly complex onboard systems.  New human–AI and AI–AI teaming approaches will be investigated with the intention of revolutionising decision making and mission planning in future military platforms – to include aircraft and land vehicles as well as ships.

In April 2019 the Defence Secretary announced the formation of a new joint military and industry technology accelerator programme, NavyX.  A total of £75 million was to be allocated jointly for NavyX’s work and for the purchase of two autonomous mine-hunting vessels for use in the Persian Gulf.   NavyX, described by the Ministry of Defence as “the Royal Navy’s new autonomy and lethality accelerator”, is intended to drive a rapid, continual transformational change in autonomy across all maritime environments, working closely with the DASA programme and providing end user testing and trialling facilities for new equipment with the Navy and its contractors.  NavyX’s first funding competition sought proposals for mature autonomy ideas for increasing the efficiency of maritime military operations and providing an improved operational capability.  The RAF has a similar innovation unit, RAFX, which works on experimental digital technology.

Following 2016’s Unmanned Warrior exercise, which tested a range of unmanned maritime systems, the Navy has also hosted an annual series of ‘Information Warrior’ artificial intelligence, information exploitation, and cyber warfare exercises.  As part of the 2019 Information Warrior exercise a virtual maritime task group was set up in Portsdown Technology Park to simulate information systems which are under development now and are expected to enter service with the Navy over the next few years.  The exercise tested automated methods for finding data to support military operations, and then processing it and moving it to the relevant personnel using artificial intelligence applications trained with data from historical datasets.

On the ground

The Army, too, is taking forward a number of initiatives which will allow it to increase its use of autonomous systems on the battlefield.  As with the other two forces, projects for pioneering manned-unmanned teaming systems are of particular interest to the Army.  DASA has funded a programme for developing manned-unmanned teaming systems for the Army, focusing on semi-autonomous forward reconnaissance systems.  In September 2019 funding of £3 million was shared by six companies – Leonardo, General Dynamics UK, QinetiQ, Horiba Mira, SCISYS and Tekever – to develop concept demonstrators controlled from manned armoured vehicles and helicopters for the Army.

Another area of interest to the Army is logistics and the automated delivery of supplies to the battlefield.  The Defence Science and Technology Laboratory (DSTL) has been working for the past three years to test the use of driverless vehicles and drones for delivering supplies to combat troops in the field.    In September 2019, DSTL collaborated with the US Army’s Combat Capabilities Development Command to test and experiment with prototype semi-autonomous logistic convoys and ground and aerial autonomous resupply systems at the Camp Grayling Joint Manoeuvre Training Centre in Michigan in the USA.  As well as demonstrating the use of delivery convoys, the exercise also trialled logistic planning tools and robotic and semi-autonomous loading vehicles.  The Ministry of Defence’s announcement for the exercise explains how such capabilities would add to the Army’s warfighting capabilities by “allow[ing] these missions to take place with fewer soldiers exposed, resulting in fewer casualties and freeing up troops to join the fight and increase the firepower”.  Shortly after the exercise two contracts together worth around £5 million were awarded to Horiba Mira and QinetiQ to deliver unmanned ground vehicles and enabling autonomous systems for further trials with the Army, and in a further contract four multifunctional cargo-carrying autonomous vehicles were ordered from Rheinmetall.

Unmanned resupply vehicles were one of a range of technologies tested during ‘Autonomous Warrior 2018’ – a four-week military exercise on Salisbury Plain billed as “the biggest military robot exercise in British history”.  The Autonomous Warrior exercise provided an opportunity for the Army and its industry equipment suppliers to evaluate how robotic technologies would operate in a combat environment, and develop an understanding of the abilities and limitations of the systems tested.  Over 70 different unmanned systems for surveillance, long-range and precision targeting, enhanced mobility and re-supply of forces, urban warfare and situational awareness were trialled by personnel from the Army, US Army, Royal Marines, RAF and DSTL.

One of the developments from the exercise which showed promise for the future was the ability to retrofit a standard vehicle for remote or semi-autonomous use with a control kit which can be set up within a few hours.  All in all, Autonomous Warrior was judged to be a success.  Information provided to Drone Wars by the Ministry of Defence following a Freedom of Information request showed that the experiment had demonstrated that all seven of the hypotheses under test were correct, although further experimentation was considered necessary in three areas.

Click image to open full booklet obtained via FoI

Following the exercise, the programme was brought together with other Army experimental programmes under the auspices of the Army Warfighting Experiment and funding of £183 million over three years from the Defence Transformation Fund was announced for four new robotic and autonomous systems for the Army –  mini drones, remote fighting vehicles, autonomous logistics, and robotic platoon vehicles.  Further exercises were scheduled over future years to investigate manned–unmanned teaming and the use of autonomous systems in decision making and communication.

Multi-domain systems

The DASA programme has also awarded funding through an ‘Autonomy in harsh environments’ competition for cross-domain technologies aimed at improving the performance and extending the use of unmanned systems in challenging operating situations, such as poor weather (high winds or heavy rain or snow), dense vegetation, high-obstacle environments inside buildings or caves, and contested electromagnetic environments.  Twenty-one project proposals were awarded funding of £2.1 million through the fund.

DASA has also supported a variety of other projects to investigate the application of machine learning and artificial intelligence in areas such as military logistics support, modelling and simulation, data management and analytics, and electromagnetic operations.  These include money for Montvieux Ltd to develop a ‘Prediction Toolset’ AI tool which uses current and historical information to predict the change of control on the ground between opposing fighting groups, and a machine learning analytics network to generate data for training object detection AI systems.  Funding was also awarded for a tool to predict component failure developed by decisonLabs and an intelligent text-structuring tool to assist human authors.

The use of artificial intelligence to predict and counter cyber-attacks is another area of interest, and DASA has recently announced that it will provide £1 million of backing to RiskAware Ltd, decisionLab, and Montvieux Ltd for work in this field.

UK seeking to draw upon autonomous decision-making mechanisms and sensors to destroy hostile drones

The DASA scheme also funds research into the development and application of new sensors, including advanced vision, novel radar concepts, and sensors suitable for mounting on space-based platforms.  One of the aims of the programme is to engage with technology companies of all sizes to exploit innovative approaches from across the sector, and to develop a pool of new, non-traditional suppliers from the technology sector.  Companies which have proved particularly adept at winning funding from DASA include Montvieux Ltd (artificial intelligence systems), Blue Bear Systems Research Ltd (drones and swarming), Roke Manor Research Ltd (artificial intelligence), and Horiba Mira (autonomous vehicles).

Looking to the future, autonomous systems are likely to feature in the government’s Integrated Review of Security, Defence, Development and Foreign Policy (currently on hold as a result of the coronavirus pandemic).  One of the goals of the review is to “determine the capabilities we need for the next decade and beyond to pursue our objectives and address the risks and threats we face”, including how to better use technology and data for this purpose.  Linked to the Integrated Review is a review of the UK’s defence and security industrial strategy, which among other things is intended to address the procurement of innovative technologies in future.

International arms control negotiations

On the international scene, the government continues to play a hesitant role in relation to arms control initiatives aimed at checking the development and use of autonomous weapons systems.  The UK has been a little more willing to engage on this issue in recent months and its recent statements at the United Nations contain some useful elements which could contribute to the development of a global control regime for such weapons.

On the positive side, the UK places a strong emphasis on human control over weapons systems, but on the other hand it also takes the view that an international legal instrument which would ban autonomous weapon systems would be premature and that an instrument of this kind could potentially lead to a watering down of existing international humanitarian law which governs the conduct of war.  The government also argues that a legal instrument could impede developments that could be useful militarily and to help protect civilians.

Claims that the UK has no intention of developing killer robots need to be treated with caution.  The evidence presented above indicates clearly that the government sees artificial intelligence and autonomous systems as indispensable for the military of the future, and many of the systems under development could be weaponised and enabled to operate beyond human control with relatively little difficulty.

In summary, the UK’s support for measures which would control and prevent the development and use of lethal autonomous weapon systems is lukewarm, to say the least, while it is pressing ahead strongly with efforts to gain an advantage and expertise in developing its own military autonomous systems, short of lethal weapons.  There is a lack of political courage to show international leadership in this area, and public and Parliamentary pressure will be necessary to force the government to change its position on autonomous lethal weapon systems.

REGISTER NOW

By Peter Burt
(Source: dronewars.net; May 1, 2020; https://wp.me/pXeuZ-3fV)
Back to INF

Loading please wait...