How Next-Generation Edge Computers Will Shape Drone Warfare

September 1, 2024

Drones have played a key role in defense and security operations for some time now, but it’s only over the last few years – and particularly with the devastating conflict in Ukraine – that the true combat potential for drones has been realized. The use of drones in combat is now the norm, and combined with advanced embedded computers capable of deploying artificial intelligence, is creating a capability that will give forces a superior advantage on the future battlefield.

Drones – or uncrewed aerial systems (UAS), as they are known by the military – have become ubiquitous in most armed forces around the world. Across all domains, drones employ a range of sensors and have been used traditionally in a passive intelligence, surveillance and reconnaissance (ISR) capacity, collecting vital information on enemy activity and feeding the intelligence cycle.

The advantages of military UAS – which come in a range of sizes – over traditional manned platforms are well documented, and include the following: Removing personnel from danger, longer platform endurance, reduced platform costs, larger fleets of smaller platforms and reduced risk of escalation if shot down.

As well as ISR functions, drones such as the MQ-9 Reaper have also been able to close the OODA loop (a four-stage decision making framework consisting of Observe, Orient, Decide, and Act) by carrying out lethal strikes using onboard precision-guided missiles. However, these were largely the purview of advanced militaries, until the rise of small drones.

The Growing Use of Weaponized Drones

The last decade has seen a surge in the use of small and inexpensive UAS – including quadcopters and fixed-wing examples – that can carry out lethal attacks. The first notable example of this was during the fight against Islamic State (IS) in Syria and Iraq in the mid-2010s.

IS pioneered the use of cheap off-the-shelf quadcopters that could drop hand grenades, as well as hobby RC planes that could carry explosives. While primitive, the armed drones were effective in stopping advancing forces that were attempting to retake IS-controlled cities such as Mosul.

The war in Ukraine has only served to accelerate and mature the technology and tactics of drone warfare, with much of this conflict now characterized by the use of commercial-off-the-shelf (COTS) drones for both surveillance and lethal strikes. In early 2024, Ukraine’s Minister of Digital Transformation stated that Ukrainian drones had destroyed or damaged 73 tanks, 10 howitzers, and 369 Russian personnel, all in just one week.

Ukrainian forces have pioneered the use of cheap first-person-view (FPV) drones – initially developed for recreational activities such as filmmaking and drone racing – that are configured with explosives and flown directly into a target using skilled operators. A typical mission could see one drone deployed by an ISR team, which then passes off targeting data to a separate FPV team to complete the ‘kill chain’.

Drones are now highly integrated into Ukraine’s force structure with the Ukrainian President announcing in February a new strategic branch of the armed forces known as Unmanned Systems Forces.

Future Technology Developments for Drones – the AI Revolution

As computer processing advances further and artificial intelligence (AI) is deployed on more systems at the edge, including those pioneered by Systel, then the autonomous capabilities of drones will only advance further. One example of how autonomy and AI will underpin new drone tactics is swarming, where multiple UAS will work collaboratively together to perform a mission with limited human input.

A flavor of the devastating impact of swarming drone attacks has already been seen with Russia’s bombardment of civilians in Ukraine, as well as Houthi attacks in the Red Sea and the organization’s use of Iranian-made loitering munitions on naval and merchant shipping. These attacks show how multiple relatively cheap drones can overwhelm air defenses and quickly deplete defensive ammunition stocks.

In the future, swarming drone attacks will be more intelligent and integrate various electronic warfare countermeasures that spoof and degrade air defense systems further.

The use of AI will also allow drones to be more survivable in degraded environments and, like swarming, carry out missions with limited human input. Currently, many UAS still rely on control input from a human operator and a ground control station – particularly the cheap commercial FPV drones used in Ukraine – which is challenging in environments where jamming exists and signals can be blocked.

An answer to this has been to leverage AI – including machine vision – to aid with autonomous navigation and target selection, especially in the terminal attack phase. This has been seen with Ukraine’s attacks on Russian infrastructure using long-range drones, as well as a new generation of FPV drone that features an automated terminal attack phase that overcomes the jamming of control and data links.

This autonomous target detection and recognition is likely to become increasingly sophisticated as AI-enabled edge computers become more powerful and are utilized more to process UAV sensor data. There will also be a greater degree of cooperation between autonomous systems, allowing them to work together and make decisions that will help them achieve desired outcomes, adapting as the mission requires.

The mission sets for UAVs are also likely to grow further, with much focus now on heavier-class UAVs that can be used in a logistics capacity, ferrying supplies such as ammunition, food and medicine to frontline units and freeing up crewed helicopter assets to perform other tasks instead. These UAVs will also likely be highly autonomous and have limited human input thanks to edge computing.

US Army Investments in The New Drone Era

The new reality of drone conflict has not gone unnoticed by the U.S. military, including the U.S. Army, which announced this year that it would “rebalance” its aviation modernization investments to include the phasing out of drone systems that are not survivable in modern conflicts, including the Shadow and Raven UAS.

The Army has said that it will increase investments in “cutting-edge, effective, capable and survivable” unmanned aerial reconnaissance capabilities, as well as procurement of commercial UAS. “We are learning from the battlefield – especially in Ukraine – that aerial reconnaissance has fundamentally changed,” said the Chief of Staff of the Army, General Randy George.

He added: “Sensors and weapons mounted on a variety of unmanned systems and in space are more ubiquitous, further reaching, and more inexpensive than ever before.”

As part of its rebalance, the Army will accelerate efforts on the Future Tactical Unmanned Aircraft System (FTUAS), a program that Systel, a leading rugged computer manufacturer based out of Sugar Land, TX, is supporting and one that will give Army Brigade Combat Teams a rapidly deployable ISR capability. Another key program for the U.S. Army is Launched Effects (LE), which will see small UAS launched from Black Hawk helicopters to provide key data for commanders during large-scale combat operations.

As the U.S. Army progresses with LE, it is also likely that the service will begin to test swarming capabilities and how one operator can control multiple UAS from one device.

Computing Requirement Grows for Drone Warfare

As drones incorporate more intelligent features, particularly AI-enabled ISR and targeting capabilities – including target detection and recognition, pattern of life (POL) and behavior recognition, and predictive analysis – as well as more autonomous functionality including swarming, the requirement for more powerful edge computing that can capture and process data “on-prem” at the edge will grow.

It will no longer be viable for data – including time-sensitive intelligence data – to be sent back to a command HQ for analysis, which can delay the decision-making process and potentially lead to mission failure. Equally, autonomous decision-making at the platform level will also remove the need for constant human-in-the-loop operations.

Edge computing will enable this, reducing the reliance on centralized HQs or control stations, which are vulnerable in contested environments and where bandwidth is limited or not available. This allows tactical commanders or platforms themselves to make speedier decisions, often in real-time, that will translate into an operational advantage.

As edge computers will have to operate on the front line, they will have to be rugged enough to withstand the harsh environments that UAVs will experience, as well as reliable enough to survive the high tempo of major combat operations. This will mean that they must meet and often exceed military-grade environmental standards such as MIL-STD-810, which covers ruggedization specifications including temperature, vibration, shock, altitude, humidity, and sand and water ingress protection, and MIL-STD-461, which covers electromagnetic interference (EMI) and compatibility (EMC) ruggedization specifications.

Additionally, in order to maximize interoperability and upgradeability as well as cost efficiencies, edge computers need to be engineered with open standards using a Modular Open Systems Approach (MOSA).

Edge computers also need to be designed around the limited size, weight and power (SWaP) available on UAS platforms, which has been an engineering challenge for drone manufacturers ever since their inception. Advanced AI processing has traditionally required larger processors and more power, and this has been difficult to achieve on power-constrained platforms such as UAS, where space is also extremely tight.

The Latest Edge Computers

The latest generation of military edge computers are now capable of providing significant performance – especially in data-intensive applications such as edge AI and processing – in smaller and smaller form factors. This is particularly important for the UAS space, where SWaP considerations are critical and the requirement for AI-enabled systems and autonomy is growing.

The latest in MIL-SPEC rugged embedded computers leverage the very latest in commercial-off-the-shelf technologies (COTS) and are capable of fusing sensor capture/encode, processing, networking, storage, controls, and distribution into a single powerful and highly configurable Line Replaceable Unit (LRU).

These computers also leverage system-on-modules (SOMs), which incorporate key components onto a single printed circuit board, which are ideal for embedded and edge applications.

SOMs can integrate GPUs, CPUs, and memory, and are particularly suited for tasks that require significant processing power, such as video analytics that leverage machine learning algorithms, as well as allowing multi-sensor platforms such as UAVs to analyze information at the source. Higher-end SOMs can carry out very complex AI and robotics tasks, although this will require greater demands in terms of cost and SWaP.

These components are also leveraging next-generation interfaces that ensure high-speed and efficient data transfer, which is critical as the amount of data produced through multiple UAV sensors increases. These interfaces include PCI Express, now into its 5th generation (PCIe 5.0), which Systel is rolling out as more components become compatible.

This article was written by Aneesh Kothari, President, Systel, Inc. (Sugar Land, TX) For more information, visit here  .

---