In the driving seat

The hunger for automated vehicles (AVs) – by the civilian and military sectors alike – has been a constant for decades. One expert, Bryan Clark, senior fellow at the Hudson Institute, a US-based thinktank, has looked back at the ways in which AVs have been a goal of automobile manufacturers for decades. However, they “have failed to reach maturity due to the inherent complexity of passenger driving conditions, which span everything from well-defined highways to chaotic urban streets and unmarked country roads.

“Driver-assist technologies, however,” Clark adds, “have made dramatic improvements in vehicle safety, essentially breaking down the automated driving mission into a set of discrete tasks that are assigned to the vehicle or the driver depending on who is best suited. Managing speed, direction and lane alignment are tasks automation does well. Making decisions on whether to leave a standing stop or turn are often best left to the driver, as some well-publicised accidents showed over the past several years.”

The US Department of Defense (DoD) and allied governments, both inside and outside Nato, are focused on procuring ‘human-in-the-loop’ and ‘leader-follower’ type automation systems – the former requires people to be involved in an automated system’s decision-making process, while the latter sees a human-driven vehicle followed by an autonomous vehicle at a set distance. The advantage of this is that development and evolution of commercial truck automation is proceeding rapidly, due in part to the infusion of massive venture-capital financing over the past six years. Within the DoD, many in the top brass see the importance of moving forward with deployment of human-in-the-loop or leaderfollower automation in order to bring automation more rapidly to key parts of military operations, including military supply convoys as well as various types of combat convoy applications. Expanded deployment of automated convoy or leader- follower-capable vehicles can also be complementary to the longer process of developing automated individual vehicles that can operate on their own in various types of environments.

A great big convoy

Trucking is an AV use case that allows more of the tasks involved with driving to be shifted to the vehicle. Shipping depots and distribution centres generally have well-defined traffic patterns and are close to major highways with minimal complex city traffic in between. As a result, human operators can offload nearly all the tasks involved in driving to the vehicle’s control system and focus on supervising. The US Army has been aggressively studying military vehicle convoys. Unlike their civilian counterparts, such convoys “may have to go anywhere, preventing automation from relying on known routes, road markings and traffic rules”, notes Clark. “They also often must travel through areas lacking a formal road network. The US military has been addressing this using automation technology that allows a lead vehicle operated by drivers to guide automated vehicles along the route.”

This reduces the number of human drivers needed and reduces the risk to soldiers by only placing them in the lead vehicle. To allow communications between vehicles that may be separated by obstructions like buildings or terrain, “the army is using drones as communication relays”, Clark says. “In many of these experiments, automation in the lead vehicle allows the human driver to focus mostly on decisions regarding where to go, while the truck manages the tasks of avoiding obstacles and proposing routes using Google Maps-like decision aids.”

The British Army has committed to enhancing its operational logistic capability through the use of Robotic Autonomous Systems (RAS) and is exploring the best ways to do that – including experimentation with its allies – according to retired Major General Simon Hutchings OBE, master general of logistics with the Royal Logistics Corps and formerly director of joint support at the UK Ministry of Defence.

“This includes optionally crewed leader-followercapable large goods vehicles to enable our logistic personnel to increase delivery throughput,” Hutchings notes. “We see operational advantage in enabling greater endurance at increasingly longer distances with greater volumes of materiel, enhancing operational effectiveness. The RAS ambition also includes smaller autonomous systems, such as drones, which can be optimised to undertake the last-mile logistic tasks – for example, distributing blood product and critical spares.

“All of these require integration of sensors to enable data capture and usage at a standard and volume that has yet to occur on the battlefield,” he adds. “Understanding where and what is required, to maximise the benefit of RAS is critical to harnessing both volume and speed that they can offer.”

However, Hutchings is aware that the logistics liability for operating these systems must be fully understood. “Maintaining a resilient supply chain for autonomous systems will be essential – as will the ever-closer integration of our industry partners. It is also clear that a conversation about how skills may be transferred into the military workforce if needed in a crisis must be had, else we will fail to ensure that autonomous systems can be sustained and repaired in a conflict zone.”

Machine-human cooperation

According to Sagie Evbenata, senior research analyst at Guidehouse, “The ongoing mobility transition toward automated, autonomous and zero-emission vehicles (ZEVs) is transforming how the global commercial and defence communities move people and goods – the traditional boundaries of how vehicles are fuelled, stored, and utilised are being redefined.” Furthermore, a changing climate and an increase in extreme weather conditions are creating an urgent need for companies and militaries to adapt mission capabilities to incorporate resilient equipment and installations.

On the defence side, local commanders will need to assess the impact of incorporating these new vehicle capabilities and needs into their daily operations. Evbenata notes that “these vehicles will require updates to existing infrastructure and operating procedures, specifically in the areas of fleets, fuels, and facilities. These updates will also be critical to enabling new vehicles to achieve their mission while creating a more resilient infrastructure network.”

The British Army’s RAS ambition is seen as essential to maintaining a leading and modern army composed of teams made up of both humans and machines. “A key objective is the minimisation of risk to human life by the increased deployment of automated air and ground systems,” says Evbenata. “As a result, the army is involved in a number of research projects to assess and develop automated technologies with military potential.”

Evbenata cites one example of this, Project Theseus, which “looks to identify self-driving ground and air technologies to provide last-mile deliveries of military supplies such as ammunition, food, and other critical items to the battlefield”. As part of this project, the army has been developing and evaluating uncrewed ground vehicles (UGVs) that are capable of transporting cargo over off-road terrain in high-risk environments. Evbenata notes that “the army has partnered with Rheinmetall to retrofit AI-powered automation systems to some of its fleet of Polaris MRZR-D4 light strike and reconnaissance all-terrain vehicles”.

Furthermore, the army has been testing Multi-Utility Tactical Transport (MUTT) robotic platoon vehicles that, in addition to transporting infantry equipment, provide intelligence, surveillance, target acquisition and reconnaissance. MUTTs can be configured to travel semi-autonomously or by remote tele-operation.

A veteran of the 2003 invasion of Iraq, Michael Noonan, senior fellow at the US-based Foreign Policy Research Institute, argues that “as technologies mature, the automation of trucks is of importance to both civilian industry and the military. While some of the uses of this technology would overlap – increasing the safe delivery of supplies and materiel over extended distances – there would also be large differences in their uses. Neither civilians nor the military would completely drive the innovation and diffusion of technological advances in this space.”

Noonan thinks it is also important that we don’t think about any automation of supply trucks as simply replacing humans with unmanned systems. “This process, particularly in military settings, would almost certainly be the pairing of man and machine to swiftly, accurately and safely deliver supplies across territory where movement might be impeded or harassed by enemy forces, criminal elements, or displaced persons,” he says. “The ability of crew to operate the vehicles as necessary, defend and protect the vehicle convoys, and perform necessary maintenance will be essential. One could also think about scenarios where a human operated system could control other vehicles remotely.”

Automating military trucks is not simply a matter of pure research and development or science fiction conjecture. In 2011, for instance, the British Army deployed the Squad Mission Support System vehicle, also known as the ‘Ox’, to Afghanistan. Noonan saw that “these remote-operated vehicles were able to carry out missions such as delivering ammunition to special operations forces in harm’s way or construction materials to units in rough terrain”.

If one was to speculate about pure automation, however, one can readily imagine certain scenarios where that might be beneficial. Noonan’s view on this is straightforward: “One would be delivering supplies across particularly hostile territories where losses to artillery barrages or precision strike systems could reduce risks. Another would be delivering supplies across an area that was contaminated by nuclear, chemical, biological or radiological hazards – although decontamination of the systems would still be necessary. Lastly, one can imagine scenarios of military deception where convoys of vehicles would be sent in directions meant to deceive the reconnaissance, surveillance and target acquisition capabilities of enemy forces.”

A long road ahead

Regarding the path ahead for AVs, energy and transportation consultant Steve Boyd says that “there remains a long road to commercially profitable and scalable solo driverless automated vehicles. In the meantime, new generation advanced driver-assistance systems, leader-follower systems and other connected vehicle solutions can be commercially deployed more widely. And this process is well underway, despite overall macroeconomic headwinds. This type of human-inthe- loop solutions also is complementary to the longer process of developing commercially viable individual driverless vehicles.”

Although some AV companies continue touting nearterm commercial deployment plans, Boyd’s expectation is that it may take tens of billions of dollars in further investment for each major driverless AV company seeking to deploy complete systems. And it will require a decade or more to see fully driverless individual vehicles become commercially profitable and scalable on open roadways and in non-controlled environments.

It is clear that a reckoning is underway in the market, one which will be especially harsh for AV companies that fail to develop significant interim revenue products or alternative long-term sources of capital. For those that are able to survive and adapt, partial automation products or related solutions can also be important stepping stones on the much longer road to profitable fully driverless systems. Boyd argues that, “Over recent years, many companies have begun to develop and deploy interim products that can be deployed nearer term – such as advanced driver assistance systems (ADAS), safety systems and automation systems that enhance the safety of human drivers, systems that require extensive direct supervision by remote human operators, and automation systems only for use in highly controlled environments – such as truck/trailer yards, private mining/logging roads, controlled routes between warehouses, mining sites, construction sites and farm fields.”

Leaders at the DoD see a long timeline to the deployment of fully driverless vehicles. As Boyd notes, “They are focused on the practicality and safety of human-in-the-loop automation systems like vehicles with a mixture of leader-follower, ADAS and individual automated vehicle capabilities.” Further deployments of these types of military systems, largely supported by OEMs and suppliers with affiliated commercial businesses, will complement the commercial development and deployment of similar solutions.