US Army Plans Biggest-Ever Experiment with Robotic Tanks
In 2022, the US Army intends to test an entire company of unmanned combat vehicles in a simulation of a real military conflict, which has already been called an unprecedented and major step toward refining the hardware and software that will one day allow wheeled robots to enter the battlefield.
Ross Coffman, a US Army general and director of the Next Generation Combat Vehicle multifunctional combat team, told AUSA (Association of the US Army) reporters that the closest precedent to the upcoming army robot exercises at Fort Hood, Texas, was last year's exercises of military platoons at Fort Carson, Colorado. For these exercises, the army converted several old M113 armored personnel carriers into robots.
“Just imagine: if you can convert an M113 armored carrier, you can turn almost anything into a robot. We learned a ton of new information. There were clear winners in the technology base. Some vehicles were not as good. Now we are moving to the company level. The lessons learned here can be applied to a brigade and division, and we can see how we want to tackle some of the challenges in the future. But I don't know of any country that has conducted the single vehicle experiments mentioned above”

In summer 2021, the army hired five companies to develop prototypes of combat unmanned vehicles: Point Blank Enterprises, Oshkosh Defense, BAE Systems Land and Armaments, General Dynamics Land Systems and American Rheinmetall Vehicles. As part of the deal, the combat vehicles must conform to an open systems architecture that can include third-party software and sensors.
Program managers hope this will let them use the country's best minds in vehicle autonomy, machine vision and artificial intelligence. In recent years there has been an explosive growth of private sector and academic research in driverless driving and autonomous and partially autonomous programming. But the best minds in autonomous driving are not necessarily best equipped with hardware to build combat vehicles that meet military specifications.
So the army has already developed what it calls the core of robotic technology, a so-called code base for future operating systems. They will use it to test autonomous software from various sources.

“Essentially, this allows us to compete with modular robots and get the best devices in their class from industry, academic or government partners, where we can compete on the level of behavior. We can integrate the core into the end element that merges with these systems. So we are working through this process. We are now on the verge of selecting six suppliers who will continue to improve this autonomous software for competing behavior models”
What is this behavior? And how much autonomy does the army ultimately seek in unmanned combat vehicles? It all depends on which AI solutions come from technology companies and what the army believes is ultimately achievable. But an approach based on open systems architecture means that as new AI capabilities emerge in labs, the army will not have to do much intensive software development to integrate them into the vehicles.
“The government controls the interfaces, standards and the way data moves through the system. But we don't need to control individual black boxes, processors or sensors. We just need to control how they 'communicate' with each other. So we are not buying intellectual property we don't need, but we have the ability to remove a modular component and replace it with another without changing the architecture and redesigning the entire system”

This should allow the army to ride the wave of increasingly advanced artificial intelligence and autonomous systems products that will emerge from Silicon Valley and academia in the coming years, without relying on software that will be outdated by the time it is deployed.
Hardware and software will first be brought together in a simulated environment to test how well everything works together before the army moves on to costly assembly. This is a similar approach to what the army is now doing to prepare for its large "Connect All" experiment, which will take place later in autumn 2021.
“First and foremost, we have a systems integration lab that combines both software virtual environments and hardware representing these systems. And we will do everything in a virtual world that simulates the experiment in the systems integration lab, and then apply it to real vehicle systems to see how everything planned works. This is the interactive approach we are applying”