News

2020-02-09

A Mind of their Own: Drone Autonomy and Security

In recent years, unmanned vehicles (UVs) have received worldwide attention both as high-impact military technology and immensely popular toys for hobbyists. More commonly referred to as drones, advances in robotics technology have also elevated commercially available UVs to prominence as powerful business tools.

With the emergence of artificial intelligence (AI) and machine learning, drones will soon incorporate more autonomy and require less human guidance. As machines become more programmable, it may become possible for one individual to unleash multiple machines that carry out a pre-programmed mission on their own, while AI may even offer these machines the ability to think like humans and make corresponding decisions, including those with potentially lethal results.

At the same time, terrorists and other threatening non-state actors have also realised the benefits of unmanned technology to deploy drones in paramilitary attacks, presenting a technological reality that security managers are increasingly likely to face in the coming years. This article focuses on the short- and long-term impacts that AI and autonomy will have upon drone threats.

Emerging Drone Threat

On 4th August 2018, two commercially available aerial drones, each equipped with a kilogram of powerful plastic explosives, were used in an attempted assassination against Venezuela’s President Nicolas Maduro. Although the machines were unsuccessful in reaching their target, the attack raised the alarm and brought drone threats to the forefront of security conversations.

Drone-based attacks from the air, land and sea are not a new phenomenon, but recent technological advances and the proliferation of commercially available drone technology have simply made UVs a high-tech tool in the arsenal for these threat actors. Terrorist and paramilitary actors have shown both the capability and intent to adjust attack platforms and enable the use of this technology to increase the overall impact or lethality of attacks.

Drone technology can be categorised in several ways, although there are two key distinctions. First, drones can be classified by their availability, with the key division between over-the-counter products commercially available and military-grade products.

There are also various types of drone platforms suited for different operating environments. While the term “drone” is most often used to refer to unmanned aerial vehicles (UAVs), there are also drones designed to operate on land, the sea surface and underwater, known as unmanned ground vehicles (UGVs), unmanned surface vehicles (USVs) and unmanned underwater vehicles (UUVs) respectively.

Most of the focus in the drone industry surrounds UAV innovation as the drones are, unfortunately, most used in terrorism, both as reconnaissance and attack platforms. UAV innovation has been largely fuelled by the smartphone revolution with many such components – gyroscopes, accelerometers, GPS, processors and cameras – integrated into the onboard technology for modern, commercially-available UAVs.

Key technological advances in the UAV industry have included changes to the size of aerial drones, their payload capacity and their flight longevity. However, the most promising (and concerning) advancements in UAV technology, and the drone industry as a whole, are in the area of autonomy.

Terrorist Drone Use

As over-the-counter drone technology has evolved, non-state actors have sought to leverage the technology to carry out surveillance or attacks, with several groups having been particularly successful here, including ISIS. In January 2017, ISIS announced the establishment of a drone unit known as “Unmanned Aircraft of the Mujahideen,” which organises the group’s unmanned aircraft campaigns on the battlefield.

According to ISIS documents uncovered, their drone programme is not a series of one-off incidents, but an institutionalised, bureaucratic unit planning drone weaponisation since 2015. ISIS has also released propaganda footage illustrating the group’s ability to drop munitions onto crowds and to hit stationary vehicles and tanks.

U.S. Drone Spending

Military applications for drones have also increased dramatically in recent years. In the 1990s, the U.S. military spent around US$300 million per year on drones, but annual military spending on drones has since increased to over 23 times pre-9/11 levels, from large drones like the MQ-9 Reaper to hand-launched models like RQ-11 Raven.

For Fiscal Year (FY) 2020, the Department of Defence has requested approximately US$3.7 billion in new spending on unmanned/autonomous systems technology and US$927 million for artificial intelligence and machine learning projects. According to Phil Finnegan, co-author of the Teal Group report, 2018/2019 World Military Unmanned Aerial Systems: Market Profile and Forecast and Director of Corporate Analysis at the Teal Group:

“Worldwide military UAV production is expected to exceed US$90 billion from 2018 to 2027, with US$40.4 billion coming from Pentagon procurement. Global UAV production is projected to more than double from $4.9 billion in 2018 to US$10.7 billion in 2027, a nine per cent compound annual growth rate. Pentagon procurement spending is expected to grow from US$1.6. billion to US$3.9 billion annually, with a total research-and-development investment of US$21.5 billion over that period.”

Although the personal and commercial market for drones has primarily focused on UAVs, military usage continues to grow in significance across land and sea platforms because non-state actors have also proven an ability to acquire and utilise multiple types of military drones in operations against both military and civilian targets. For example, Houthi rebels in Yemen unveiled their drone capabilities in January 2017 not in the air but at sea, by using an armed, unmanned maritime craft to strike a Saudi warship in the Red Sea, killing two sailors and injuring three others.

Around the same time, Houthi rebels also displayed their UAV capabilities and in September 2019, they attacked two key oil installations inside Saudi Arabia, damaging facilities that process the vast majority of the country’s crude output and disrupting world oil supplies. Twenty-five military drones and missiles were used in the attack that forced the Kingdom to shut down half of its oil production.

Yemeni Houthi rebels claimed responsibility, but most people think it could not have been achieved without the aid of Iran, being that the drones themselves were almost certainly of Iranian origin. Iran is known to sell its military drones to allies and other sympathetic countries because Hamas and Hezbollah have also employed military-grade, Iranian-made UAVs.

Autonomy and Drone Swarms

While most consumer and military-grade drones are controlled by a human operator, new developments in UVs have allowed these machines semi-autonomy through components enabling the machines to interact with their surroundings. Some over-the-counter UAV models now have the ability to avoid obstacles and follow pre-determined paths autonomously during the flight. If these AI trends progress, then drones will develop the ability to ‘think’ for themselves and carry out assignments free of human control.

Concerns are now mounting over drone swarms of multiple, autonomous drones acting in unison to overwhelm a target, either choreographed in advance or controlled by multiple humans. Pre-programmed drone swarms have already been employed on the battlefield, such as the attack on Russian bases in Syria in 2018, but when combined with other technology like explosives or facial recognition software, malicious drones could become a much more significant threat.

Fears about drone autonomy are the basis for Slaughterbots, a short video from the Future of Life Institute depicting a dystopian future in which militaries build autonomous, explosive micro-drones that fall into the hands of terrorists. This sensationalised dramatisation may overlook several key assumptions, but the technological components depicted in the video are real and there is little to be done to keep this technology out of the hands of terrorists.

Drones and Physical Security

Many countries are testing autonomous drones operating with manned aircraft, like the U.S. Air Force’s unmanned "Loyal Wingman" aircraft and the Broad Area Maritime Surveillance (BAMS) system of Poseidon P-8 maritime patrol aircraft and unmanned TRITON aircraft. Further development of unmanned systems continue to progress, which will see them dispatched from manned aircraft in order to work independently or in extension of the ‘mother aircraft’, such as the recently tested PERDIX nano drones, of which 100 drones were dropped from a F-18 ‘mother aircraft’.

The ‘PERDIX’ is a micro-drone swarm system developed for the US DoD/Naval Air Systems Command together with MIT Lincoln Laboratory (WarLeaks). Swarm technology is entirely dependent on autonomous processing and these many developments or aspirations are well described in, for example, the U.S. planning document USAF RPA Vector - Vision and Enabling Concepts 2013-2038.

The prospects of autonomous technology, be they flying drones, underwater vehicles or other lethal weapon systems, clearly bring new opportunities for military forces, but if nefarious actors are certain to obtain and use autonomous drone technology what can be done now to prepare to counter this threat? For security managers in the private sector, the most effective approach to preparing for offensive UV use would be to focus on defensive measures integrated into an overall security management strategy, which would include conducting risk and vulnerability assessments, as well as integrating countermeasures into security protocols and policies.

Security managers should consider setting specific rules or guidelines to manage drone encounters and determine how the threat should be reported, what emergency response measures could be implemented and how security personnel should respond. Drills could also ensure that personnel know how to respond in a crisis involving a drone threat, while security managers should consider evaluating current technologies able to detect and disable drones.

In recent years, most of the innovation focusing on counter-UAV technologies employable for both commercial and military use, is largely based on jamming radio frequency and GPS signals that drones use to communicate with their human controller or determine their flight path. However, as drones become more autonomous, these countermeasures may become less effective.

Ethical Perspectives on Autonomous Drones

Legal regulations and compliance are also critical because many offensive countermeasures designed to disable UAV threats are not legal (domestically and in many foreign countries), due to anti-jamming regulations or those meant to protect civil aircraft.

The relevance of ethics in debates on autonomous drones must be recognised. Compliance with the law is essential to any military and political policymaking, including the development and use of autonomous drones. Although law and ethics could overlap, there may be important ethical issues to consider, particularly in the case of emerging military technologies, that have at present not been properly addressed by law. Therefore, ethical reflection could complement the law by providing guidance in such “grey areas”. It may also be important in emphasising when ethical obligations should exceed legal duties in the interest of good political governance.

The delegation of life-and-death decisions to nonhuman agents is a recurring concern of those who oppose autonomous weapons systems. One of the most primary concerns is of allowing a machine to “decide” to kill a human being, which undermines the value of human life. To make moral judgments about who may be targeted in the “fog of war” is difficult even for human soldiers. The fear is that allowing autonomous drones to make such distinctions could result in civilian casualties and collateral damage. Even if such weapon systems would be able to discriminate between combatants and non-combatants, would they be able to assess whether an attack is proportionate or not?

On the other hand, autonomous drones would be able to process more incoming sensory information than human soldiers and could therefore make well-informed decisions. And since the judgments of machines would not be clouded by emotions, it could potentially reduce the risk of war crimes that may otherwise have been committed.

Deploying autonomous drones could certainly improve aspects of humanitarian missions, benefiting civilians who are being assisted and reducing risks to soldiers. It could be used to search dangerous areas or perform high-risk tasks, such as bomb disposal, and reduce the risk of human soldiers being injured or killed.

However, there are also debates about the fact that limiting the risk to soldiers by removing them from the battlefield could make war too “easy” and might even reduce it to a low-cost technological enterprise that no longer requires public or moral commitment.

The irony is that autonomous machines without centralised command-and-control systems present a unique type of threat where brute-force offensive countermeasures may be the most effective approach. The legal environment for drone countermeasures will be just as important to follow as technological developments.

Deciding Future Directions

The technological potential of autonomous drones is already being tested and developed but the extent to which they will become important military technologies will depend on national requirements, which will be determined in turn by the future security situation. It would be better to develop a legal and ethical framework before we come to such a situation because autonomous drones raise important judicial and ethical issues about responsibility for unintentional harm and create some gaps in moral accountability.

When autonomous military systems are deployed it becomes less clear how to apportion responsibility, yet such potential responsibility gaps must be addressed through technical solutions and legal regulations, requiring the engagement of NATO and allies in international discussions on these topics. Hence, defending against drone threats will require security managers to take both an immediate and long-term approach, being that the threat already exists but is likely to evolve in the coming years.

The reality is that UVs of various shapes, sizes and capabilities are already employed by hobbyists, businesses, militaries and threat actors. As the capabilities of drone technology increase and the barriers to use decrease (e.g. cost, accessibility, operability, etc.), terrorists and other nefarious actors, including lone wolf attackers, will become increasingly likely to adopt the technology.

AI and autonomy will likely enhance this trajectory of upward use and enable more lethal, coordinated attacks with less human involvement. Although threat actors have not yet achieved the ‘killer drone’ swarm capabilities predicted in the Slaughterbot film, their less autonomous predecessors have already arrived and should be addressed as part of a comprehensive drone countermeasure strategy.

Anti-drone Solutions

While AI may create new, frightening opportunities for threat actors, it could also provide a drone security solution. Some of the anti-drone measures used by military at the moment are the following:

SRC Inc’s Silent Archer: As a counter-UAS system-of-systems, Silent Archer includes options for both fixed-site emplacement and on-the-move shipboard operations with radar, EW and camera to detect, identify and defeat small UAS threats flying individually or in swarms. SRC’s counter-UAS technology is currently deployed by the U.S. Air Force and U.S. Army.

Lockheed Martin’s ATHENA: Lockheed Martin has recently demonstrated their laser weapon system for the U.S. Air Force at the Fort Sill, Oklahoma government test range, where the system successfully engaged and shot down multiple fixed wing and rotary drones.

The Advanced Test High Energy Asset (ATHENA) operated in a fully netted engagement environment with a government command and control (C2) system and radar sensor. The radar track was provided to airmen who operated ATHENA via cues from the C2, then ATHENA's beam director slewed, acquired, tracked and defeated the drone with a high-energy laser.

Validating this type of full kill-chain performance has been a priority of the U.S. Air Force and other branches of the Department of Defence. It remains a requirement for laser weapons to be effective against unmanned aerial systems (UAS) on the battlefield.

ATHENA was operated by USAF personnel during this demonstration, where it was able to destroy multiple drones in engagements representative of what is being encountered by U.S. armed forces. As the ATHENA high-energy laser system is transportable, it enables the Air Force to emplace it wherever they need to defend bases and high-value assets.

Coyote UAS and KURFS: Equipped with an advanced seeker and warhead, the Coyote-enabled radar system can successfully identify and eliminate threat UAVs when paired with an advanced electronically scanned array KuRFS radar, which acquires and accurately tracks every size of UAS threat.

Deployed from the ground, air or a ship, Raytheon’s Coyote is small, expendable and tube launched. Coyotes can be flown individually or netted together in swarms, being adaptable for a variety of missions including surveillance, electronic warfare and strike.

In addition to Department of Defence missions, the versatile Coyote is also used by the National Oceanic and Atmospheric Administration for hurricane tracking and modelling. The U.S Army is conscious that advanced Coyote variants will fly faster and farther, having decided to use Coyote as a counter-UAS solution since 2018.

Today’s KuRFS radar is a multi-mission radar providing rocket, artillery and mortar, sense-and-warn counter-UAS mission capabilities, with a level of accuracy enabling significant UAS mission performance at tactically significant distances, including precision fire control and UAS swarm scenarios. More than 40 KuRFS radars have been deployed by the Army, who are upgrading the systems for extending capability and ensuring support beyond 2025.

“Think” Like a Human

A central question for the multiple drone detection and neutralisation platforms available concerns their requirement of a high level of technical competence on the part of the operator. Such expertise is hard to find outside of the military.

Training the system to “think” like a human could thus allow security personnel to achieve a high level of detection without human involvement while limiting false alarms. AI may also provide a number of security benefits by augmenting human capacity in other areas of physical security like conducting perimeter patrols.

Reference Text/Photo: www.hsdl.org,www.lockheedmartin.com,www.raytheon.com, www. srcinc.com

Når dronene våkner: Autonome våpensystemer og robotisering av krig” (Oslo; CappelenDamm, 2016)

Comments () Views (3408 ) Print

Send To Friend