Uncrewed Ground Systems

Potential uses include carrying cargo, casualty evacuation, reconnaissance, chemical-agent detection, communications and fire support. However, the gap between ideal uses and present technical capability is significant. The delivery of systems to where they will be used, the realistic uses once there and the machines’ interactions with soldiers have frequently been underexamined but are crucial to how UGS will form part of land forces and offer genuine operational advantage. The technical limitations of UGS must be reflected in how they are task-organised within land forces. Due consideration must be given to how UGS will move around the battlefield, as it will often not be by their own steam. Maintenance and repair of UGS will require new training courses and a close relationship with industrial partners.

The principal conclusion to draw is that UGS will require significant support from their human counterparts. Moreover, cognitive burden on operators must be considered and managed. Systems move slowly, and the difficulty of navigating in complex terrain means they are not suited to some of the tasks for which they have been proposed, such as dismounted close combat in complex terrain. It is important to involve as many soldiers as possible in experimentation, and expose them to UGS early and often. This can be achieved by employing UGS in those areas with the highest throughput of soldiers, such as firing ranges and exercise areas, and making use of simulation. In addition, initial training should include education and demonstrations of UGS for new recruits. This will help build familiarity, favourability and trust in these systems.

The potential of human–machine teams is significant, but hype should not disguise the limitations of UGS and the difficulty of integrating new technology into established structures.

Introduction

Context

The presence of robots on the battlefield is central in today’s military discourse. A recent British Army recruiting advert showed soldiers operating in close combat alongside humanoid and wheeled robots. A former head of the British armed forces has stated that in the 2030s, the Army could comprise 90,000 soldiers and 30,000 robots. The Chairman of the US Joint Chiefs of Staff said in a recent interview that “you’re going to see significant portions of armies and navies and air forces that will be robotic”. There is a significant jump from where forces are now to this envisioned state. Experimentation with uncrewed ground systems (UGS) in military forces is gaining pace. Many forces are running trials with a variety of systems. Uncrewed aerial systems (UAS) are far more mature in their journey and lessons can be drawn across to their land-based brethren. Similar to UAS, it is believed that UGS will provide competitive advantage to land forces in several ways. UGS have the potential to support logistics and reconnaissance missions, as well as the ability to be armed with remote weapon systems to provide additional firepower to manoeuvre units. They can remove soldiers from harm’s way and increase mass, which underpins fighting power. However, there are substantial technological hurdles and organisational realities which need to be overcome before UGS are seamlessly integrated into military forces and become a force multiplier. The simple existence of such systems is not enough to transform warfare or generate competitive advantage for a force. It is not clear that any military force has integrated UGS at scale except for bomb disposal robots. These basic UGS have been part of military arsenals for decades, but the current zeitgeist is focused on those systems with a degree of autonomy which can unlock operational effectiveness above that seen on battlefields today.

This paper answers three research questions focused on the integration of UGS into light land forces at the tactical level. The first concerns how UGS can be usefully employed in tactical land formations with their technical limitations and tactical realities considered. The second relates to how they get to the fight in the first place: organisation, movement and sustainment of UGS around battlefield echelons must be considered, and this is much less examined in the literature than is their use in frontline combat. The third involves how military forces can ensure that UGS are put to good use by their soldiers. Preparing soldiers to form part of human–machine teams must be a deliberate act, using training and education to build trust and understanding. The paper focuses on developments in the British and US militaries, but lessons can be drawn more widely.

Light land forces have been chosen as the focus for discussion, although employment considerations can be extrapolated to other parts of the force. Light infantry operate with minimal vehicular support, although they may be supported by vehicles such as quad bikes. They have the critical task of closing with the enemy at close quarters and seizing ground in complex terrain. These troops are laden with all the equipment required to operate for days at a time, including weapons, ammunition, rations, water, radios, batteries and more. As a result, they may have much to gain from the advent of UGS.

This variety of potential uses means that UGS offer great potential utility to armed forces. However, their development, introduction and scaling across armies requires careful consideration, the totality of which is not immediately obvious. Considerations are set out below to outline how military procurement professionals and concept developers might conceive the introduction of UGS into the force.

Structure

After setting out its methodology, the paper outlines the principal characteristics of UGS. These are the basis of their numerous uses and the foundation of their strengths and limitations. The drivers behind UGS development – including reducing risk, increasing mass, and the ability to increase advantage through human–machine teams – are noted. Next, the state of the art of UGS is shown, demonstrating the numerous use cases which are developing in forces around the world. With this foundation set out, the bulk of the paper then offers several areas of investigation and recommendations for military forces. The first concerns how such systems might be employed at the tactical level. The second is how UGS can be moved around the battlefield and where they might be assigned organisationally. Third, means by which to socialise UGS within a force, improve education and foster trust are offered. These areas are often sidelined by discussion of experiments or capabilities, without due thought to the various interdependencies and whole-force considerations.

Methodology

This paper is founded on both primary and secondary research. First, the author has conducted consultations with both practitioners and analysts, aimed at discussing their experience with UGS and associated technology. He has also deployed on and visited military exercises, such as Project Convergence 22. The author is a serving military officer and has extensive experience of and a professional background in the employment of robotics and autonomous systems (RAS). He has spent time with industry, looking at both hardware and software. A literature review of academic articles, news media and military press releases has also been conducted.

The author has also attended conferences with military personnel examining UGS. Existing research and expertise on the organisational impact of UGS is limited compared with that on their aerial counterparts. The literature is either very technical with an academic focus, or less analytical, mainly comprising news articles and manufacturer comment. Moreover, the paucity of information in the public domain about military UGS has also imposed a limitation on this research. Attempting to describe a future state is inherently difficult, but the assumptions and considerations laid out in this paper are grounded in reality, and draw on practical knowledge of both RAS and military organisational processes and structure.

I. What are UGS?

UGS are vehicles or static platforms that operate on land without a human crew inside, although some systems can be optionally uncrewed. UGS can be as small as shoeboxes and even thrown by users. Others are as large as historically crewed vehicles, weighing many tonnes. They may or may not be armoured. UGS may be wheeled, tracked, have legs or a combination of the three. Each type of drivetrain has its advantages and disadvantages. Wheels are good for speed and manoeuvrability on even surfaces, are lightweight and are simple to replace. They are, however, vulnerable to shrapnel damage and punctures. Tracks are useful for offroad manoeuvrability and offer good traction over rough terrain. However, they are generally slower than wheels and are also complex to refit if they become dislodged. UGS with legs, such as the Boston Dynamics Spot, can tackle obstacles such as stairs and climb very steep slopes, and can also move laterally. Wheeled and tracked vehicles are faster over most surfaces, however.

UGS exist on a spectrum of control. They may be operated by a soldier holding a wired controller or a remote control while within line of sight. Examples include mine clearance systems and bomb disposal robots. Teleoperation adds a level of complexity, in which the operator relies on the UGS’ cameras and sensors to make sense of surroundings and controls them from a distance. UGS with levels of automaticity or automation are more complex still. Within this category, there remains significant variety. It is necessary to stress that a system being uncrewed does not mean it is autonomous. The Autonomy Levels for Unmanned Systems (ALFUS) framework is one toolset with which to understand UGS’ autonomous capability. Autonomy can be understood as a system’s “own ability of integrated sensing, perceiving, analyzing, communicating, planning, decision-making, and acting/executing, to achieve its goals as assigned”. Systems with high levels of autonomy are rare. More commonly, UGS have a leader–follower function whereby the vehicle will follow another crewed vehicle or a human commander. Increasing levels of autonomy then allow some UGS to follow waypoints given by a human operator and avoid obstacles while following a given route or exploring a designated area. Some systems may have the capability to act with conditional automation, whereby an operator can take control in certain circumstances, such as if the UGS cannot figure out how to manoeuvre around a certain obstacle. UGS that have the capability to act independently of an operator’s instructions and make a series of linked “decisions” in pursuit of an end objective are scarce. And that end objective will have been given by a human operator, which again means that the system is not fully autonomous. The necessity of human input is a golden thread in this research. Supervision of many systems still requires soldiers to be at least monitoring, and perhaps solely focused on, the UGS, rather than free to conduct other tasks.

Systems also differ by use, which is examined in detail in later chapters. For the purposes of this paper, it will be assumed that any remote weapons systems associated with UGS will have a human in the loop throughout for decision-making, retaining meaningful control, and providing authorisation for any engagement. This is in line with British defence policy.

Sensors

The simplest remotely operated UGS may have no sensors, as the human operator is expected to be close by. An example might be an excavator. Systems such as bomb disposal robots have cameras that allow the operator a close-up view from the system, and allow the manipulation of the target object with the operator at a safe distance. As systems gain autonomous functions, a suite of sensors can be expected, including LIDAR, RADAR, GPS and cameras. LIDAR and RADAR help the UGS make a 3D map of their surroundings, which is then used for routing and obstacle avoidance. Ultrasonic sensors may be mounted on the sides of the vehicle to detect objects very close up. In civilian applications, these are used to help autonomous vehicles park. Video cameras are used to detect humans or animals, as well as to make sense of traffic lights and signage. Video cameras are also able to pick up more nuances than LIDAR and RADAR, including hand gestures and traffic cones. GPS helps the system situate itself within the wider geography of the area and aids a system to stay on course when navigating a waypoint route or searching an area for reconnaissance purposes. UGS may also have an inertial measurement unit to give an additional indication of the direction and velocity in which the system is moving. This information can complement that of GPS, and is useful when GPS signals are weak, such as when moving through urban areas or tunnels, or during bad weather. Developments in this area are fast moving, and new sensors and combinations are being experimented with. Given this, commenting categorically is difficult, but suffice to say UGS use sensors to make sense of their surroundings.

Software

Software must then make sense of all the inputs described above. The UGS’ use, the environment they will operate in and their level of autonomy determine the complexity of the software. The software uses the sensors to make sense of where the UGS are, what is around them and, in some cases, what might happen next in the case of people and vehicles in close proximity, and what to do if a particular circumstance presents itself, such as another vehicle moving into the systems’ path. Software will use this information to plan UGS’ next move before moving. The systems may take their external environment and plan against a library of scenarios on which they have been previously trained. The software must fuse information from the various sensors to form one combined understanding of the environment, using a variety of filter mechanisms. Software architectures differ from system to system and in complexity. UGS may also have target recognition capability that can spot armoured vehicles and movement on the horizon, which can be fed to commanders for subsequent decisions and actions.

Power

Smaller UGS are usually battery powered, with larger systems using a combustion engine or hybrid diesel–electric power train. Each has benefits and limitations. Electric systems are near-silent to run and produce a low heat signature. However, battery life is often limited, and requires extensive management, of replacing batteries and charging them. Systems using diesel or petrol are easier to fold into existing military logistic chains as they are already geared to provide fuel to current fleets. However, they have a significant noise and heat signature, which can make them vulnerable in an era of persistent ISR capability.

Command and Control

Despite the connotations of autonomy, UGS must in practice remain connected to their human operators. This could be to give the UGS instructions on where to go next, or to execute a particular command. Or it might be to relay information back to the operator, such as a potential target. Data processing may take place at the edge, depending on the size of the platform, or packets of data will be sent for processing elsewhere. UGS will place demands on the existing combat radio network, and this must be planned for. There also exists opportunity for adversary action in jamming or spoofing systems. UGS may be able to carry out tasks without being connected to the network, before reconnecting when necessary, which will increase their survivability.

II. What are the Purported Benefits of UGS to Tactical Land Forces?

The drivers for the development of military UGS are numerous, and are broken down below.

Risk

Using uncrewed systems in place of crewed vehicles can reduce risk to personnel. Soldiers can be kept further back from the line of contact and can avoid a number of dull and dangerous tasks that up to now have been the responsibility of humans.

Mass

Uncrewed systems allow the generation of additional mass above that which can be formed through an army’s physical workforce size. A future scenario might see one soldier controlling a suite of UGS, which could increase the area over which a unit has sight, influence and, potentially, control.

Situational Awareness

UGS equipped with sensors such as cameras and radar can help commanders get a firmer sense of the battlespace. UAS have proven very effective in this area, and UGS can add additional capabilities, such as navigating through those places less accessible to UAS.

Burden Reduction

UGS can carry equipment that currently burdens soldiers. This allows soldiers to move more quickly and with less effort. This is important when soldiers have become loaded with equipment – in the pursuit of protection, reducing their ability to fight.

Human–Machine Teaming (HMT)

In the popular imagination, machines replace people in their roles entirely. However, this is not how military forces are conceiving of the near to medium horizon. Instead, the optimum balance between soldier and robot is key. HMT makes use of the comparative advantages inherent to humans and machines respectively. Humans do the tasks they are best suited to, and robots do those they are best at. The British Army, for example, envisages that humans will remain the core part of HMT for some time to come. The Army framework sees increasing machine involvement over time. In the immediate future, RAS-enhanced teams will see machines used in a transactional manner, as tools. These teams are limited by the current levels of autonomy and human levels of trust. This phase sees machines used to increase performance in human-led tasks. Later, trust and technology develop to enable RAS-integrated teams in which humans cede more control to machines whose autonomous capabilities have improved. Here, humans and machines perform tasks that result in a combined outcome. Finally, RAS-supervised teams are envisaged in which machines can outperform humans and humans retain a supervisory role to keep meaningful control.

This framework is particular to the British experience, but a similar gradient is noted in other forces. For example, the US Army’s RAS strategy notes three likely epochs of development. The first lasted from 2017 to 2020, when the Army matured concepts and initiated programmes to look at increasing situational awareness, lightening the load on soldiers and improving sustainment. The second epoch, from 2021 to 2030, aims at improvements including achieving automated convoy operations and removing soldiers from lead vehicles. In the far term, from 2031 to 2040, the first era of automated systems will be replaced, and see new organisational designs and fully integrated autonomous systems, which work in concert to achieve the task.

It is not the case that simply adding systems is the answer to providing mass in armed forces. Depending on levels of autonomy and the requirement of a task, soldiers can only manage so many responsibilities. If an uncrewed ground system is remote controlled without any level of autonomy, the ratio will be one to one, or even worse. It has been noted on some experiments that it takes three soldiers to adequately manage one uncrewed ground system. A one to two ratio would see one soldier jump between systems to operate them. Systems with more autonomy are less burdensome on the operator, and soldiers can then manage more systems at once. Cognitive overload is a crucial consideration when building a force structure that includes UGS. There are only so many screens or notifications a soldier can make sense of. There are also more practical considerations that do not generally make it into discussions of HMT at the policy level. On Project Convergence 22, a US military experimentation exercise, a US Army officer spoke of the difficulty for a junior soldier of sitting in the back of a moving Bradley armoured fighting vehicle while trying to manage uncrewed systems on a tablet computer. They quickly became overwhelmed. This might be because the uncrewed systems required inputs or verification from an operator, or it might be because the information and intelligence being sent from the systems was difficult to digest. Simply sitting in an armoured vehicle on the move is not a comfortable experience. Adding additional cognitive load may be problematic. Ergonomic issues such as motion sickness are an important consideration. Some soldiers may cope better than others. Seemingly minor additional tasks may have significant repercussions for combat effectiveness. This speaks to the importance of allowing soldiers to get used to working with such systems, and being aware of their own abilities and those of the systems.

Having outlined the foundational concepts of military UGS, the potential individual tasks of such systems can be investigated, the subject of the next chapter.

III. What are the Potential Uses of UGS?

UGS have several proposed uses for military forces, some of which are more obvious than others. These are identified here as potential uses, while subsequent chapters tackle the realities of their employment, whether such uses are realistic, and the implications for the force.

Load Carriage

Figure 1: UGS with Cargo Basket

Load carriage is the principal identified task for UGS at today’s stage of development. This might be carrying personal equipment such as bergens, rations and ammunition, or platoon and company equipment such as ladders or beaching equipment. UGS might also be equipped with stretchers to enable casualties to be extracted from danger areas. Casualty evacuations are a particularly strenuous activity for soldiers. Being able to use UGS instead has multiple benefits. It allows soldiers to preserve energy in close combat, where fatigue can lead to poor decisions and further casualties. It also keeps soldiers free to complete the task at hand, such as winning a firefight. Another related use for UGS is for broader logistic purposes, especially in the dangerous “last mile” delivering supplies to frontline locations.

Communication Node

Figure 2: UGS Fitted with Radio Equipment

UGS could carry a unit’s radios, which can be very heavy and slow to move. They may also carry electronic countermeasure and electronic warfare systems, which can be used to prevent explosive devices detonating, or to disable enemy UAS. Equally, there are times when soldiers must be detached to form a rebroadcasting or retransmission service if radio waves are blocked by terrain or another barrier. This allows units and headquarters to communicate with one another. This task might be completed by a UGS with a communications equipment fit.

Surveillance and Reconnaissance

Figure 3: UGS Fitted with Cameras and Sensors for Surveillance and Reconnaissance

UGS can be equipped with sensors that can scan the area for potential threats. Software can categorise objects in the UGS’ field of view and identify points of interest, both static and mobile. These can then be passed to commanders for further investigation and potential targeting. Another use of UGS is as a reconnaissance screen moving ahead of dismounted or mounted recce soldiers. Or they might be employed in a static or roving function around unit locations or bases.

Chemical, Biological, Radiological and/or Nuclear (CBRN) Sensing

Figure 4: UGS Fitted with CBRN Sensors

UGS can provide a sensor capability for CBRN threats. UGS with appropriate sensors could be sent to locations of potential attacks. Equally, they could remain with troops and carry sensor equipment that had previously to be carried by soldiers.

Armed

Figure 5: UGS with Remote Weapon Station

UGS can be armed with remote weapon stations. Remote weapons are in mainstream use on crewed armoured vehicles today. Their benefit is that they allow the weapon to be fired by operators from inside the vehicle without a soldier having to be exposed in a cupola. Cameras mounted on the system allow the operator to aim the system and maintain control. Such systems, for example the Kongsberg Protector, can be mounted on UGS and operated remotely by offset troops. Such weapons might be used as sentry devices or in a fire-support capacity. Another potential use for UGS is as mobile landmines, a technique that has been adopted by the Ukrainian armed forces fighting Russia.

Engineering

Figure 6: UGS Fitted with Mine Clearing Capability

Military engineering includes the breaching of obstacles, demining and providing plant for trench digging. This is currently done by hand, or by soldiers using excavators. The civilian mining industry is a world leader in uncrewed technology and uncrewed diggers are in common use. UGS with a digging capability could set up a defensive position with much less human input than is currently required.

Deception

Figure 7: UGS Fitted with Emitters for Deception

UGS might also be employed to provide deception capability. This could be in the form of “fake” vehicles or groupings, or they can be used for deception using the electromagnetic spectrum. Such systems deliberately radiate to mislead the enemy. UGS equipped with a radio system and antennae can be used to draw enemy resource and disguise intentions and dispositions.

UGS may be multirole and capable of carrying out more than one of these tasks at a time, or of switching between them. Moreover, UGS should not be considered in isolation. There are also UGS built as mobile launch pads for UAS, such as the THeMIS Observe, which is an example of using the two technologies in concert. Military strategy requires conducting the orchestra of military capability in the most suitable way possible. UGS should be used for those tasks where they offer a competitive advantage. They should not be the answer before the question has been asked. There is always a danger of pursuing technological innovation for its own sake, especially in times when commitments outstrip resource – which is a place in which many forces find themselves. This friction has been recognised as problematic in military forces in the past, and has at times resulted in poor decisions.

Having introduced UGS and their proposed military uses, this paper moves in the next chapter to answer three questions:

• How can UGS realistically be employed today and in the immediate future, with technological limitations and tactical realities taken into consideration?

• How are UGS task-organised and how do they move around the battlespace?

• What is the best way to ensure that soldiers use UGS as intended?

The soldier must remain central to these efforts. The uses outlined above broadly represent attempts to do away with human input where possible. However, UGS are built to support soldiers in their endeavours, and it is soldiers who will enable them to do this. The relationship is key, and the focus should remain on the human, as demonstrated below.

IV. Considerations for UGS Support to Light Manoeuvre Forces

Gently Does it

UGS lack manoeuvrability in close or complex terrain. This must be a central consideration for their employment in tactical formations. Their ability to troubleshoot when faced with obstacles is currently far below that of humans. When moving autonomously, UGS must make sense of their surroundings to plot a clear path. Navigating obstacles using sensors alone is incredibly difficult. A study using the TAERO optionally crewed wheeled system found that “it is possible to effectively implement autonomous mode up to a speed of 2.8 m/s in an unstructured environment”. Advertised maximum speeds for UGS far exceed that which would be possible in complex terrain. This pattern is seen in numerous trials and reports, in which soldiers outpace their robotic counterparts. This finding is further corroborated by wargames and testing. The civilian transport sector is yet to make autonomous vehicles a viable offering despite billions of dollars and years of research and development. This is also in spite of a relatively robust framework within which they must work. Road networks have defined edges, junctions and rules. The latter are not always followed, of course, and autonomous vehicles on roads must try to account for the actions of other road users, which cannot always be predicted. The problem becomes more difficult when extrapolated to military UGS. Normal road networks are a simpler environment than a battlefield, where smoke, debris, adversarial activity, and disturbed earth make for a much more complex picture, with fewer established norms. Water hazards are illustrative here. Water’s surface is highly refracted, meaning it looks different depending on the view angle, the surrounding area and the weather. In wet weather, determining what is simply a slick surface versus a puddle versus something deeper is difficult for sensors and computers.

The vision of autonomous land systems moving around the battlefield with abandon is currently fantasy. Most systems that are advertised as, or considered to be, autonomous or AI-enabled are much more limited in their capacities. As noted above, uncrewed does not mean autonomous. For example, the Milrem Robotics THeMIS is one of the more advanced and developed platforms on the market, with buy-in from several European countries. It can be teleoperated and can complete waypoint navigation as given by an operator. At the time of writing, a “follow the leader” capability is still in development, as is the ability to swarm. Teleoperation is usually conducted using a line-of-sight antenna. As such it is limited by terrain and range. In the case of the THeMIS, the line-of-sight range for control is up to 1,500 metres. This central limitation is clarified when overlaid with the proposed tasks of UGS outlined above.

Combat

Figure 8: UGS as Fire Support

Dismounted close combat is an inherently complex business. It involves rapid decisions, movement, adaptation to constantly changing dynamics, and the most intimate of command and control, communication and logistic interactions. As a result, such activity will remain the realm of humans. UGS are far from being able to close with and kill the enemy on an objective. There are simply too many variables for systems to manage coherently, and the systems’ vulnerabilities too many.

However, AI-enabled systems can add value by accurately sensing and categorising objects in their field of view, providing important information to the commander. Sensors and their respective algorithms can distinguish between types of vehicles, military and civilian, with great accuracy. One study showed a 97.25% to 99.5% detection rate at 2,000–5,000 metres, both during the day and at night. Another, using different methods, achieved accuracy of above 85%. The fact that these systems are not achieving 100% accuracy is not a reason for alarm. People are fallible and contend with issues of eyesight, optics, climate and fatigue when engaging in combat. For UGS, these figures will only improve with time and access to labelled datasets, which will in turn grow as the proliferation of UGS continues.

In the current state of development, armed UGS are probably better placed to provide supporting fires. This task would traditionally be done with a fire support section set off to a flank while another section carried out the assault. Supporting deliberate offensive action lends itself to the use of UGS, as the terrain can be analysed by commanders ahead of time. In this scenario, armed UGS are likely less suited to ad hoc offensive action and instead must be used deliberately. The idea of robots facing off against other robots while humans sit in a command bunker watching the action unfold is misleading. Placing three armed UGS in a fire support position with a human in the loop for engagement authority, and soldiers adhering to battlespace management boundaries, is a more realistic application, balancing well understood norms with novel technology. Equally, static defence tasks such as an anti-tank screen might be envisaged. This matches UGS’ and soldiers’ relative strengths.

Supply

Figure 9: UGS for Supply

Resupply is one of the more mature tasks for UGS, and this is one where most experimentation has been completed. At the larger scale, platoons of uncrewed heavy goods vehicles might be led by a crewed lead vehicle for logistic missions in rear areas. The logistic and movement constraints outlined mean that the use of UGS in rear areas is the place to focus attention. However, due to risk to personnel, current research focuses on autonomous “last mile” resupply. In fact, rear areas are also now vulnerable, in the face of persistent ISR and precision strike. There is, therefore, value in fielding UGS in these areas, where tasks and wayfinding are often more simple than using main supply routes. Fielding UGS here would also allow data collection, which is crucial for system improvement.

Currently, it is likely that a human would still be involved in these tasks, providing a lead element to be followed, either on foot or in a crewed vehicle. However, UGS would still be useful, as logistic patrols are a significant burden on forces. Reducing crew requirements to free up soldiers to do other tasks is an important contribution of UGS. The urban environment provides an avenue through which UGS could be employed further forward, as moving between buildings leaves soldiers vulnerable.

That said, a slow-moving UGS would be an easy target for enemy troops. There is a tension at the heart of the proposed use of UGS for burden carriage in combat scenarios. The dismounted troops who have the most to gain from having a system carry their equipment are also those who need to be able to move rapidly through complex terrain such as forests and urban environments. Smaller vehicles may be more agile, but they cannot carry that much equipment. While UGS could reduce what soldiers are carrying, they would add friction if they were unable to keep up in tactical movement in complex terrain due to technical limitations. There may be scope for these systems to follow units a tactical bound behind, but there is a risk that they could get stuck. This then becomes an additional constraint and planning consideration for commanders. Therefore, it is sensible for UGS to remain with companies or the battlegroup echelons, where movement will be more deliberate.

Reconnaissance

Figure 10: UGS in “Stay-Behind” Reconnaissance Function

Employing UGS in a reconnaissance capacity would see lines of robots moving in front of the traditional human recce screen. At present, soldiers move ahead of the formation’s main body looking to spot the enemy before the enemy spots them. This enables shaping activity and for deliberate targeting by indirect fires to take place, which is preferable to having to react on someone else’s terms. Recce is also risky. Recce units are generally small, detached from the larger mass of their formation and susceptible to interdiction by the enemy, which is in turn looking to achieve the same effect in reverse.

A concept proposed in the US supports deploying a forward line of RAS, thereby reducing risk to personnel. A forward line of sensors can probe positions for enemy activity, and potentially force them to unmask. This could be by moving and giving off a signature, be it heat or electromagnetic, or by engaging the UGS, which also gives away their position. However, the limitations discussed above demonstrate that this vision is a long way off for UGS. The use of UGS in this way would slow manoeuvre units to a crawl, making them susceptible to targeting from enemy fires. In addition, there would be significant burden in trying to manage their movement and make sense of their data. This task is best left to UAS. UGS with this function are best suited to static, and perhaps predesignated, roving sentry tasks, where they can support soldiers to maintain situational awareness over an area. A situation where UGS could be used as a “stay behind” capability as friendly troops withdraw is a more suitable use case, and more palatable than using soldiers in what is a very risky activity. Leaving UGS to identify the movement of enemy troops and vehicles and alert friendly forces plays to their strengths in image recognition. It also has the advantage of freeing up recce troops for additional tasks.

The considerations for deployment in these three areas can be mapped across to the other potential tasks outlined earlier in the paper. Those tasks that require high levels of mobility remain under the purview of UAS. CBRN threat monitoring and radio rebroadcasting can be achieved by UAS, although there may be times when UGS are better suited to the tactical situation. This chapter has considered the technological limitations associated with various types of UGS, and has applied these to tactical formations. The next chapter looks at the enabling activities needed to ensure that UGS are in the right place in working order.

V. How Do UGS Get to, and Stay in, the Fight?

Military logistics have been brought into sharp relief by the war in Ukraine. The true potential of UGS can only be unlocked if they are in the right place at the right time for the right task. Like other military equipment, UGS will need to be transported to the area of operations. The size and ability of the system will determine how this might happen. Factoring UGS into future lift capability, on land, at sea and in the air, is important for planners. Military lift capacity is a limiting factor to the success of deployments. Every system that is transported takes up space that cannot be used by another piece of equipment. The military benefit in theatre must therefore be clear. Units and formations are responsible for devising field equipment tables for the kit they need in theatre to do their job while deployed. UGS will feature in these considerations going forward. There is little capacity for superfluous equipment. Larger armoured systems such as the Milrem Type-X, a 12-tonne uncrewed system equipped with 50-mm cannon to support main battle tanks, or the 10-tonne General Dynamics TRX, will need dedicated logistic support. Larger vehicles are moved by aircraft or low-loader trucks and ferries. In the near term, all these options require human crew, emphasising the reliance of UGS on people. Smaller systems such as the Milrem THeMIS, which is the size of a small car, can be towed behind a parent vehicle until they are required. That parent vehicle will need to meet specific towing requirements, such as height of hitch. In the case of the THeMIS, the speed at which it can be towed is three times as fast as it can move itself – 80 km per hour, rather than 20 km per hour. Moving UGS from an initial railhead, port or airfield to the area in which they will be employed must be planned for in detail.

The totality of the system must be considered, including power supply. If the UGS are battery powered, how and where are these batteries charged, and who does the charging? Which echelon should be burdened with the charging capability? Battery technology is relatively nascent, and stamina remains low. On battery power, the THeMIS has a runtime of just one and a half hours. In hybrid mode, using its diesel engine, it has a runtime of 15 hours. Low-level battery management for existing equipment such as radios already requires planning and demands electricity, which may be provided by the mains, generators or other vehicles.

Another consideration for UGS is where repair and battery charging take place. In the case of crewed vehicles, the crew can fix small errors and conduct simple repair jobs. For instance, great pride is taken by tank and artillery howitzer crews in their ability to fix a track if one becomes dislodged. UGS will not have the luxury of an on-hand repair crew. This means that resource must be dedicated to recovering systems once broken. Repair functions in military forces have become eroded in recent times, as systems have become more complex and manufacturers retain the right to repair. The ability to repair equipment and keep it on the battlefield has been shown to be crucial in the conflict in Ukraine. For instance, a third of Ukraine’s howitzers are out of service for repair at any one time. Repairing technical equipment is often left to contractors rather than completed in place, even for well-established capabilities that are in service. Sensors and computer systems, no matter the platform on which they sit, are vulnerable, despite ruggedisation by the manufacturers. Holding UGS back several bounds until they are used for a discrete task before being recovered will allow more sustained repair operations than can be offered at lower formations.

Managing demand for UGS by frontline units is another concern for planners. As in the case of UAS earlier in their development, demand for their support far outstrips UGS supply. It is still the case that larger and more capable UAS are held at divisional or corps level and assigned to discrete tasks depending on a commander’s decision. Specific recommendations for UGS are difficult to outline without firm knowledge of the types and numbers of systems to be procured. They will likely be a scarce resource for some time. However, forces should be wary of putting manoeuvre units in permanent possession of larger, more capable UGS. If soldiers are having to consider what their UGS are doing instead of fighting the enemy, then the systems have been misemployed. Tactical units should bid for UGS support as they currently do for aircraft. In this framework, bids for support from aircraft are submitted while formations are planning for future operations. The demand for aircraft for offensive support, moving people or cargo, or providing reconnaissance and surveillance, generally outstrips supply, as platforms are scarce. To that end, units make bids for capability, and a central cell determines who gets what and when. This generally works on a rolling 72-hour time horizon tied to the operational area’s planning cycles.

Currently, formations bid for a primary and secondary asset to provide support. The primary would be ideal, but may be tasked elsewhere, so a second, different asset should also be identified. In this case, with UGS in their infancy, the secondary course of action should employ established capabilities. This will mitigate against undue reliance on UGS while the capability is nascent.

Network

It is not just the physical systems that need to be in place. UGS with a reconnaissance or surveillance function need to be able to relay that information back to commanders, using a robust communications network. That network may also need to permit some UGS to pass information among themselves, either to corroborate a potential target if more than one system can “see” it, or to help them avoid obstacles. Equally, commanders may need to issue instructions to the UGS for a task. The electromagnetic spectrum is not an unlimited resource, and different capabilities must be deconflicted. Radars may interfere with aircraft if their systems operate within the same band. The network needs to remain available and have enough capacity to pass information around. This is the focus of major experiments, such as the Project Convergence series, in which a resilient network is identified as a “backbone” to enable large amounts of data to be passed around. This is easier said than done. Militaries use a host of different communication systems and bearers, from radios through to satellites. The network needs to have low latency, be efficient in its use of bandwidth, and be secure from enemy interference. All additional interactions with these networks provide adversaries with opportunities to interfere. They may look to jam or spoof UGS. Robust countermeasures will need to be in place, or UGS will suffer in the same way UAS have in Ukraine, with 10,000 systems lost a month. What is more, the network needs to be interoperable with those of allies and partner forces. Importantly, it is likely that the network will be provided by a different company, or set of companies, than those who have built the UGS. A variety of bearers, data links and data standards make interoperability very complex. In a contested network space, prioritisation of the information being transmitted is important.

Adversary Activity

While providing opportunity for friendly forces, the proliferation of UGS also provides options for the adversary. This might include jamming GPS or seizing control of systems using electronic warfare means. Systems with automated navigation and reconnaissance capabilities are also vulnerable to adversarial attacks on their software. Here, machine learning and AI models can be “attacked” by objects in the physical environment, where an input specifically designed by an adversary can cause a system to act in an unamenable way. An understanding of a system’s software architecture and logics can allow an adversary to confuse a system and reduce its effectiveness, or deduce the information on which it has been trained. Researchers tricked an autonomous vehicle into misidentifying a stop sign as a 45 miles per hour sign, a mistake that could have had catastrophic consequences. Subtly altered images that look normal to humans can fool AI. In one study, a 3D-printed model of a turtle was specifically designed to trick a computer into thinking it was a rifle, which it did at every angle it was presented to the camera. Such activity is worrying in relation to sensors that seek out targets in a given area, as there are rules of engagement in which possession of a rifle might allow targeting. This shows the importance of maintaining meaningful human control in such systems. Adversarial activity is also troublesome in relation to more benign UGS with logistic functions that may be convinced to stop or get trapped maliciously by adversary action.

This said, the ability for real-world adversarial attacks to be successful is limited. The complexity of defeating multiple sensors in the physical world outside a research environment is a significant barrier, and may simply make such attacks uneconomical. Some of the ability to counter adversary activity will be built into systems by developers. However, military users who are alive to the threat will be better able to manage it, which raises the importance of awareness and understanding, discussed in the next section.

Force Design

Force structures will look different as UGS become more prevalent. Maintaining the same force structure and simply adding UGS on top will not maximise advantage. One frequent claim is that robots will replace soldiers in some cases. However, it is unlikely that this will be a zero-sum relationship, in which more robots can lead to forces having fewer soldiers. The British Army is experimenting with how force structures might change via its Experimentation and Trials Group, and initiatives such as the Phalanx platoon, which has reimagined the traditional platoon structure for when more uncrewed assets are integrated. In the near to mid-term, a rebalancing of forces into support functions may be required, as the example below demonstrates.

Force Design Lessons from UAS

One must look at the whole uncrewed ecosystem to see the interdependencies and how an army with many uncrewed systems might look. The best real examples today involve UAS, as the more mature capability. The British Army’s Watchkeeper is a fixed-wing surveillance UAS. It measures six by ten metres and weighs 450 kg, requiring a runway to operate. It operates on a line of sight data link with an endurance of around 14 hours and a range of 150 km. While it has no pilot inside the aircraft, the personnel and logistic tail is significant. The aircraft is operated by two pilots in a ground control station, with a third required at times. A nuance here is that military pilots can only have an eight-hour duty period, which includes flight planning. Given this, for Watchkeeper to be used at full capacity, two or even three sets of pilots are required. Watchkeeper does not have the ability to taxi and does not have ground brakes, as a weight saving measure, increasing endurance. To this end, it employs a groundcrew of seven to ten people, depending on experience levels and instructor requirements. The groundcrew tow the aircraft to the take off point and run pre-take off computer scripts alongside the pilots in the ground control station. They also set up the cable system that is used to recover the aircraft on landing. Away from the runway sits an engineering detachment of around 20 people. It conducts routine maintenance on the aircraft and keeps it airworthy. It also constructs and dismantles the aircraft when it is loaded into shipping containers for transport. It is supported by two field service representatives from the aircraft’s manufacturer. These people provide technical support and a link back to industry, which can provide in-depth technical support when needed. In addition, a command and flight operations staff of between five and ten people manages the sorties and liaises with wider airfield stakeholders. It manages the risk profile of the aircraft’s flights and provides the wider support wrap to the soldiers in the detachment.

In this case, one uncrewed system requires a wider staff of over 40 people for it to operate in a benign environment on an established operational airfield. What is more, the infrastructure required to store, transport and maintain the aircraft is a significant footprint.

While exact roles and ratios may vary, this example is indicative of the challenge of employing uncrewed systems. While such systems technically remove soldiers from a frontline task, the tail of necessary support will likely be extensive, at least in the short to medium term. For example, the key enabler for UGS is the availability of engineers to keep systems running. New technical trades focused on computer-systems engineering will be needed. Software changes rapidly, and it is likely that the burden of keeping engineers up to date with latest developments will be considerable. In turn, this will mean new courses will need to be designed, with an important question being: who would be the right authority to design such courses? These courses will then need to be run from a base, requiring accommodation, classrooms and hangars. The integration of UGS fundamentally changes the size and shape of the force using them.

This section has made it clear that humans will be the key enabler for UGS – they will move them around the battlefield, they will fix them and they will manage them, at least in the near to medium term. Thus, while it is seemingly logical to focus on technology, it is the soldier who will unlock that technology’s potential, and indeed use it as they see fit, which will be discussed in the next chapter.

VI. How to Make Sure Soldiers Use Them

Integrating new technologies into a force is difficult and should not be considered on a solely technical basis. Scaling the use of UGS across a land force is a deliberate organisational change programme. This chapter examines the role of experimentation, training and trust on the route to successful HMT. Actual future users of UGS, not the abstractions of experimentation, must be front and centre in these endeavours.

It is a mistake to assume that soldiers use equipment given to them in the way intended by designers. One trial saw soldiers continually overload a UGS, as its capacity was not enough for their needs. This led to the system overheating. At the other end of the scale, it should not be assumed that soldiers will use UGS at all. A host of factors interact to determine how soldiers use the kit they are issued. These might include previous experience, who trained them and when they were trained. One example here is personal load-carrying equipment. The British Army brought in a new type of body armour and load-carrying equipment – Virtus. However, many soldiers opted to keep using their old equipment, as it better suited their purposes. They could carry all their equipment, they knew where everything went, and it had worked so far in their career.

Experimentation

Experimentation is important for understanding the utility of new capability. New technologies are generally examined and researched for a broad use case. Then they will be handed over to troops for a pilot programme, before potentially being rolled out more widely. For all the talk of the importance of such technology in future warfighting, there is little evidence that forces have started to integrate UGS on a regular and even basis. Many soldiers are not being exposed to uncrewed technologies, even if forces may think they are. UGS integration is vulnerable to becoming stuck in an experimental purgatory, on a small scale that disenfranchises the rest of the force. An order from the Dutch Army Command to a single officer was to “just get started and explore the possibilities” of RAS. While an admirable aspiration, this is too tentative. Experimentation often takes place with a limited audience for practical reasons of scale. However, this small scale can have a deleterious effect on the success of the experiment. US Major General James Dubik refers to this increase in scale as “expanding the experimental ground”. Simulation may offer one route to democratising the experimentation process. Bohemia Interactive’s “virtual battlespace” simulation software, in use with the British military, has integrated several of the UGS discussed in this paper, for example the THeMIS. Terminals are widely available throughout the British defence estate and accessible to troops, should they be given the time to make use of them. With simulation, there is less reliance on access to physical systems, of which there are not many. Simulations allow soldiers to test approaches and witness the strengths and weaknesses of the UGS outlined above, confirming appropriate use cases. It is, however, difficult to say yet how this will impact the integration of UGS into the force, or actual future use.

Another difficulty in experimentation and novel procurement is the military’s propensity to replace like with like. As a result of this propensity, force structures look very similar to how they did 50 years ago. There is difficulty in identifying truly disruptive innovations because they do not look like what the organisation is currently doing. This limits organisations’ openness to the truly disruptive potential of UGS. Indeed, the discussion above itself adds UGS to existing structures, techniques and tactics. It may be the case that using entirely novel tactics may be the way to gain competitive advantage. This is where extensive experimentation with many members of the force should be considered. Giving soldiers the freedom to troubleshoot and use the system without preordained norms may lead to unexpected and beneficial findings.

Timelines for the introduction of UGS into land forces are tentative. The British Army’s RAS strategy uses horizons stretching out to 2035 for the integration of RAS, despite them having been part of force structures for decades already. Making use of corporate knowledge developed in the UAS world can help ease the frictions of integrating UGS. The US military’s timeline is more assured, but progress towards its ambitions is uncertain. The Project Convergence series of experiments led by the US hopes to merge capabilities between partner nations in the pursuit of effective integration and increased lethality.

Lethargy is common in military decision-making, and it is important that UGS do not fall into the trap that so often ensnares military procurement. The phenomenon whereby innovative technologies receive government funding but fail to make it into the hands of warfighters is known as the “Valley of Death”. Indeed, it appears that with AI being perceived as a potential silver bullet for many military issues, and RAS and UGS being the physical embodiment of that technology, militaries are having to hedge and spread their bets over a wide variety of initiatives. For example, the UK’s Defence and Security Accelerator has awarded more than £180 million to 1,065 different projects, an average of just £169,000 per serial. This is slightly less than the annual capitation rate of a single software engineer with the professional background and resources to develop this technology meaningfully. Increasing focus on those capabilities that show potential for the use cases described above is a potential route to success. Signalling commitment to the cause and allowing industry to plan accordingly is a key output of any RAS and UGS strategy. Indeed, the extended period of experimentation seen so far that has not led to serious expansion may in fact signal to industry to disinvest from research and development of UGS.

The buy-in of top-level leadership is also crucial to successfully instigating change in an organisation. In the case of military experimentation, there can be a propensity for general officers to only attend “distinguished visitors’ days”, which are designed specifically for show, providing an element of innovation theatre. These sessions involve orchestrated demonstrations to show best-case scenarios. They also often take place at the end of an exercise period, in which frictions and realities have been found and then solved or worked around. Multiple rehearsals take place and minute details are agreed on by the deliverers. Such opportunities give industry representatives access to senior officers, and will often be identified as a career-enhancing event for the organisers. This can lead to true frictions being masked, and often means that the generals who hold authority for novel equipment programmes do not have an accurate and holistic picture of the state of play. Moreover, the tendency of armed forces personnel to move roles every two to three years means that only a general, rather than deep, level of understanding can be achieved. In a fast-moving technological environment, this is inimical to progress.

Trust

A significant barrier to successful integration of UGS is trust. The desired human–technology relationship is often framed in terms of trust. This suggests there will always be some level of uncertainty about the workings of such systems, including UGS with some degree of autonomous function. Definitions of trust are numerous, and it is not feasible to give a full review of definitions here. One usable and well-cited definition of trust is, “the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party”. To get their full utility, soldiers must embrace these systems and trust them to complete a task. Another conception is that trust in AI-related technology is a contractual one. A system can be considered trustworthy if it can maintain the contract made with a human operator. That is, the system will carry out the given task.

Computer models that allow some level of autonomous activity are necessarily complex. There is a lack of transparency in many machine learning and AI models. When working with another soldier, it is possible to ask them why they made a decision, and person-to-person interaction is a norm with which all are familiar. This becomes more difficult with a “black box” scenario, where the decision-making process is opaque and not fully understood by the user. Trust is built slowly, but lost rapidly in the face of failure. Unless a system is fully explicable, a sceptical soldier is unable to query UGS as to why they want to act or have acted in a particular way. The military has many examples where lack of trust would cause a breakdown in operational effectiveness. The most obvious is a targeting system where a machine alerts a human operator to the potential presence of the enemy. Scepticism rather than over-trusting here is preferable, where a soldier checks the information before potentially suggesting an engagement through appropriate means. A more nuanced example would be the willingness of soldiers to load injured comrades on to UGS tasked with moving the casualties back to an aid post or hospital. The soldiers may think they could get there faster, and they might well be right. One study showed soldiers opting to manually control a UGV rather than trusting it to follow waypoints or a leader.

Many studies of autonomous systems are focused on the ethics and practice of lethal autonomous weapons systems. Moreover, this discussion is often happening between civilian commentators. There has been much less research on the importance of various design features to active-duty service people. One study found a direct friction between maintaining meaningful control and understanding on the one hand, and maintaining the increased operational tempo that uncrewed and autonomous systems are hoped to unlock, on the other. Soldiers need to be able to rapidly verify a system’s suggestions and decisions without having to work through the entire evidence body, which would render the system moot. To that end, Jai Galliott and Austin Wyatt suggest that confidence measures in observations by UGS should be accessible to soldiers. Such measures would not be infallible, because of the technical reasons and potential for adversarial action discussed above. Therefore, a secondary suggestion by respondents to the study cited above was for systems to have a means of both simply describing their planned actions and of confirming that UGS have “understood” their operator’s commands. It would be worthwhile to consult a wide user base on this issue, rather than only people who happen to be in small experimental units, which may be more by luck than judgement.

Equally, there is a fear of over-trust. Overestimating the ability of UGS will lead equally to an inefficient allocation of resources. This makes the process of integration and education throughout the force all the more important. Trust in automated systems has led to accidents in both conflict situations and commercial aviation. In Kuwait in 2003, a US Patriot detachment shot down a British Tornado, killing both pilots. The Patriot crew had acted on indicators given by the system’s computer. The best way to build trust is to develop understanding, which is the subject of the next section.

Socialisation

As UGS proliferate, it is important for as many soldiers as possible to be exposed to them early in a safe manner. This is crucial to building the trust that is a precursor to success in HMT. Familiarity breeds trust, but military forces are poor at introducing soldiers to capabilities that are not their core system. Familiarity can also build favourability, whereby soldiers and commanders are willing to lean on these capabilities when planning operations. Such favourability is not a given. The more that soldiers are exposed to UGS, in whatever guise, the better they will understand them and the more likely they are to become ambassadors. As noted above, building trust is crucial to the full integration of UGS. Importantly, it is recognised that trust will not be developed solely by developers improving software outcomes over time. Instead, most gaps in trust “won’t be solved by code but by conversation”.

This conversation might take place in several ways. The crucial step is to safely move UGS from being only in the hands of experimenters into those areas which see a large throughput of troops. These are most likely to be training establishments, both for initial training and for later tactical training. The first way is during military training and education. If military forces are not including modules on UGS in basic training, they should do so immediately. This might be as simple as a classroom discussion or presentation. Better still would be a physical demonstration using UGS. This could be a short session where a UGS’ capability is demonstrated to soldiers under training. The seemingly small act of having a trainee lie on a stretcher mounted to a UGS and travel a short distance would have manifest training benefits. As mentioned above, there is also an opportunity for simulation to play a role in widening the population of troops with exposure to UGS.

The second area for consideration would be training areas and firing ranges. Large numbers of troops who have gone through basic training pass through these facilities each year. Forces undergoing range work could integrate a serial using a UGS. This could include UGS with a remote weapon system providing overhead fire, a task currently done by soldiers. This would build trust and understanding and increase the audience exposed to such systems. Equally, many range serials involve a simulated casualty evacuation. A “casualty” will be designated by the training staff, and the soldiers will have to give first aid and use a stretcher to evacuate the soldier to a safe area. An uncrewed ground system with a stretcher could be in place on the range and used to show its utility and allow soldiers to interact with novel systems. Pitting a human team against an uncrewed ground system would begin to show soldiers and commanders where and how UGS can be most usefully employed – they do not necessarily need to learn this from an instructional leaflet produced by a faraway department. Instead, troops would be enfranchised by direct experience. These activities would also create additional data for the manufacturer about usage and failure rates.

Siloes

State defence enterprises are large organisations. They consist of tens of thousands or more personnel. There are central departments or ministries and single services, as well as research laboratories such as the Defence Science and Technology Laboratory (DSTL) and the Defence Advanced Research Projects Agency. Both the US and the UK have directorates dedicated to scanning the future and identifying concepts and capabilities that might be brought into forces. UGS are such a capability. It is not uncommon for people within defence ministries or the single services to not be aware of complementary activity that is taking place elsewhere within the organisation. This is a significant friction, and it prevents progress. In the UK, for example, DSTL, the Ministry of Defence Head Office and the Army Futures Directorate, which owns the HMT programme, all explore UGS. In addition, commercially, Defence Equipment and Support leads the procurement and delivery of UGS into the force. There is also the Experimentation and Trials Group, which leads experimentation with UGS. Moreover, there is a series of defence technology accelerators and innovation hubs. This list does not take into account the bulk of Army personnel who will become the users of UGS. These people should be the focus of UGS implementation. Within this large cohort, there will be a mixture of experience, aptitude and interest in UGS. If this community could be successfully tapped and exploited, there would be significant additional capacity to enhance the integration of UGS into land forces.

With such a wide breadth of activity, it is difficult to know who, if anyone, fully understands the totality of UGS research and development. Equally, within forces themselves, understanding of other units’ capabilities is often not well understood even when they are well established. Formations regularly organise briefing days so that staff can be informed of what is available to them during planning. Internal communications on this subject should be a central effort, to ensure coherence and a clear path to actual use, rather than a succession of experiments that remain in the trials arena.

Experimentation is important, but it should not be limited to small numbers of soldiers. Instead, exposure should be wide and varied to make use of the diversity of thought and talent available. The building of trust in robotic systems must be deliberate, through exposure early on in careers and regular, good-quality education. There must be a concerted effort to break down siloes in defence establishments so that best practice and knowledge can be better shared. The common theme is giving primacy to the future users of these systems as quickly as possible and at scale.

VII. Recommendations for UGS Integration

1. Role and management: Due to current technical limitations, UGS should be employed in standoff roles and in rear areas, where there is a dividend for their use. Treating larger UGS like aircraft whose support can be bid for will allow supply and demand to be managed, as well as keeping UGS from burdening low-level formations.

2. Force design: The extra demand UGS will place on engineers and enablers (the invisible tail) needs to be baked into force planning now. The management of UGS may, in fact, require more soldiers.

3. Logistic burden: The transport and storage of UGS, and battery management, must be planned for in detail, accepting that it cannot simply be added on to existing commitments, which would further stretch scarce resource. This will ensure the force-wide implications of new technology are catered for adequately.

4. Education: Education and training related to UGS should be implemented now, while experimentation is ongoing, rather than waiting until systems are formally brought into service. Basic training should include education on UGS now, even in a basic form, to begin to build trust and familiarity, easing the integration of UGS at scale.

5. Experimentation: UGS trials should be integrated into those areas with a significant throughput of soldiers, such as firing ranges. Moreover, it should be ensured that the totality of UGS experimentation and activity is understood by decision-makers and those conducting the experimentation, and that leaders maintain engagement with projects throughout the life cycle, rather than at the beginning and end. Clear ownership of the whole ecosystem is vital, while encouraging bottom-up engagement will create a user base ready to make best use of UGS.

Conclusion

This paper has discussed UGS and the considerations for successfully integrating these systems into military forces. It has described the physical and software components of such systems, and how they are anticipated to be used by military forces in the near and further future. Having established the state of the art, the paper discussed three questions.

First, how will UGS be used once they have been deployed? Systems with high levels of autonomous capability remain rare. Thus, most systems are remotely controlled or teleoperated from a distance. Potential benefits abound, such as enabling soldiers to stay out of harm’s way, and increasing the envelope over which they have sight and potentially control. UGS are not ready to manoeuvre in close combat, their movement is limited by the sheer number of variables, and humans retain the upper hand by some way. Equally, full autonomous navigation is possible, but systems move so slowly as to be potentially deleterious to their main functions, such as load carriage for manoeuvre troops.

Second, how will UGS get to, and stay in, the fight? Some UGS can be carried by soldiers, while others will need to be towed or transported to where they are needed. They will also then require collecting and moving onward to repair and maintenance before further use. A secondary effect of this is that UGS will have a significant logistic tail, at least in the short to medium term. This will lead to an increase in human enablers supporting UGS.

Third, how can soldiers be encouraged to make proper use of UGS? It is not a given that soldiers will adopt systems in the way originally envisaged by their designers, or even by military procurement officers and decision-makers. Familiarisation is key to building trust. If soldiers believe they can do a particular job better, they will follow that route. Given this, it is also important not to force the integration of UGS that do not add value to the HMT. Integrating UGS into basic training and those areas with a high throughput of soldiers will rapidly help socialise the use of UGS.

All these themes are interlinked and there are dependencies between them all. They must be considered by planners who have a firm view of the totality of the enterprise. Moving from experimentation to a capability integrated into field forces is no mean feat, and requires energy and direction from senior leadership. Somewhat ironically, it appears that the most sensible approach when considering the integration of uncrewed systems is to focus on the human.

Patrick Hinton is a serving regular officer in the British Army’s Royal Artillery. He has experience working with ground based air defence systems and remotely piloted air systems. He has also worked in the personnel space. Since joining the Army in 2014, his career has consisted of a number of appointments at regimental duty including Troop Command, Executive Officer, and Adjutant. He was the Chief of the General Staff’s Visiting Fellow in the Military Sciences Research Group at RUSI until the end of August 2023.