AUSA recap: Autonomy everywhere
Remember: Hardware is cool, but software is the differentiator.
Welcome back to the Nexus Newsletter. In this edition, we share some brief thoughts on AUSA, comment on the Army’s upcoming digital engineering and human-machine integration strategies, and provide our takes on recent news from the nexus of autonomy and national security.
Hardware is cool. Software is critical.
Unmanned and autonomous systems covered the floor at AUSA. We won’t lie: We’re just as smitten with slick hardware as the next person. Still, it’s important to remember that autonomy is a software problem, not a hardware problem. The coolest-looking piece of hardware is useless without a robust, validated autonomy stack under the hood.
Robust and validated means that a stack has performed to requirements under comprehensive testing across all of the various scenarios and conditions that the deployed system might encounter in the real world. That requires a proven autonomy software development infrastructure that leans heavily on both data tooling and virtual testing in simulation.
The next time that you see a slick piece of hardware, remember that the software is the primary differentiator and the real enabler of advanced capabilities - the software-defined combat vehicle/platform.
LTG Maria Gervais, Deputy Commanding General at U.S. Army Training and Doctrine Command, stopped by our booth at AUSA.
Digital engineering at the U.S. Army
Under Secretary of the Army Gabe Camarillo recently told Defense News that the Army plans to release a strategy to “establish a digital engineering environment meant to speed the pace, lower cost and reduce risk in weapons systems development.”
“There’s a lot of advantages and benefits to doing ... development in this fashion,” Camarillo said. “First and foremost, it reduces costs. Secondly, it gives you a chance to make trade-offs in terms of requirements and get a better capability whether you’re designing the next commercial car or you’re designing a future Army platform.” [...]
He noted a digital environment would also make it possible to test and evaluate performance before getting weapons out to range.
“It would have the potential to not only save you a lot of resources, it gets you a better capability,” Camarillo said. “It is the wave of the future, it is a direction we need to go in.”
The Army clearly has an understanding of the value proposition here.
Under Secretary Camarillo’s call-out of the commercial automotive industry is particularly noteworthy - virtual testing is standard in the autonomous vehicle (AV) industry, largely because validating AV safety and performance exclusively through real-world testing is impossible. A RAND study on commercial AVs, for example, revealed that an AV would need more than 11 billion miles of testing to demonstrate 20% better safety than a human driver, based on a human driver failure rate of 1.09 fatalities per 100 million miles driven. Even under aggressive testing assumptions, a fleet of 100 AVs would take more than 500 years to reach that benchmark. Commercial automotive manufacturers use virtual testing to quickly scale their testing capacity and to make their real-world testing (either on a test track or on-road) more targeted and effective.
An important starting point: A platform designed to help organizations validate software-defined systems by defining requirements and test cases, creating simulation scenarios, linking tests to requirements, executing tests in simulation, and creating analytics dashboards to understand system performance. Given the complexity of today’s software-defined systems, including the various unmanned and robotic platforms currently under development or on the modernization roadmap for the Army, traditional validation methods are inadequate. Anecdotal information is insufficient. Program decision makers must be equipped with a comprehensive understanding of validation progress on an ongoing basis to make decisions about what features to prioritize, bugs to address, etc.
This type of enterprise-wide validation platform is standard in the commercial automotive industry and is an important first step as the Army pursues digital engineering in a big way.
News we’re reading
Autonomy is everywhere in defense these days. Make sense of the latest headlines by reading key quotes from recent articles of interest, plus brief commentary from Applied Intuition’s government team, below:
Defense News | US Army developing integrated formations of robots and humans
Key quote: “We don’t really fight as individual systems, and we don’t fight as individual soldiers. We fight as formations, and so we’re looking more broadly than just the widget, but how a combination of technologies coupled with human decision-making responsibility will increase the lethality and survivability of our overall warfighting formations,” he said.
RCCTO and the rest of the team working on designing such formations will be on a learning campaign for a while, he said, but a future vision for the capability “would see a combination of formations that include both ground and air systems and the enablers to help find, fix, engage the enemy with humans in that decision cycle,” Rasch said.
“Whether they’re robots on the ground or drones in the air, they’re not going to work just as individual systems, but also [need to be] smart enough to work with other unmanned systems or manned systems,” he added.
Our take: Understanding how soldiers will team with unmanned and autonomous ground systems will be critical to the future success of the Army as a holistic fighting force. However, it is critical to recognize that the analysis of human-machine integration (HMI) is just the tip of the iceberg. Immediately below the surface are the autonomous systems themselves, which in turn are supported by the data infrastructure/pipeline. Merely focusing on HMI without the connective tissue between data and autonomous systems development risks drawing inaccurate conclusions about future requirements and how to fight with robots.
C4ISRNET | US Army mulls attack drones, ground robots for Replicator
Key quote: Army Secretary Christine Wormuth on Oct. 9 said the service has “a number of areas that would be right for Replicator,” including UAVs “of all sizes.” The service is in close coordination with the defense secretary’s office.
“Obviously UAVs, both using them as sensors and as deliverers of payloads, and also defending against them, are key on the battlefield,” she told reporters on the sidelines of the annual Association of the U.S. Army conference in Washington. “Some of the loitering munitions that we have could be candidates for Replicator. And then, finally, perhaps some of our ground robots.”
Our take: To date, much of the conversation around Replicator has focused on aerial systems - that impulse is understandable, given the prevalence of low-cost unmanned aerial systems with at least some degree of autonomy. In recent remarks unpacking Replicator, however, Deputy Secretary of Defense Hicks specifically called out ground-based systems: “Imagine fleets of ground-based ADA2 systems delivering novel logistics support, scouting ahead to keep troops safe, or securing DoD infrastructure.” We’re glad to hear that the Army is outlining its role in the Replicator initiative and is incorporating ground-based platforms into that effort.
Autonomous systems exist in various stages of development across the Army, other services, and commercial industry. As we’ve said before, if the DOD is to successfully integrate those systems into Replicator, infrastructure that enables continuous development, testing, and validation will be essential.
National Defense Magazine | AUSA News: Robotics ‘Most Promising Area’ of Army Vehicle Transformation
Key quote: “We have some opportunities to accelerate how we fight as an Army,” Brig. Gen. Michael Simmering, commandant of the United States Armor School, said during a panel discussion at the Association of the United States Army’s Annual Meeting and Exposition Oct. 11. “Number one … robotics. Robotics is probably the most promising area that we have in the armored vehicle portfolio to fundamentally change how we fight.”
Brig. Gen. Geoffrey Norman, director of the Army’s Next Generation Combat Vehicles Cross Functional Team, said the Army’s robotic vehicles “really represent opportunities for new ways of fighting, for organizing differently and bringing new capabilities to the battlefield.”
Our take: From RCV to XM30 and beyond, the Army is making significant progress in autonomous systems development. Those systems can help realize Secretary Wormuth’s vision of “always having robots, not soldiers, make first contact with the enemy.” We’re proud of our work accelerating autonomy deployment across several leading robotics programs for the U.S. Army through our category-leading software solutions and commercially-proven best practices for development and testing.
It’s worth noting that progress on autonomy can and should extend beyond programs that aim to develop and deploy fully autonomous systems. Rollovers and other non-combat accidents and mishaps were the leading cause of death for active-duty service members from 2008 to 2021. Many of these accidents could be prevented by equipping new and existing platforms with the kind of Advanced Driver Assistance (ADAS) technologies that are widely available in the commercial automotive market.
Autonomy can transform how the Army fights in a fundamental way. In parallel, it can also make seemingly-marginal improvements that, when viewed across thousands of deployed platforms, have an equally significant impact on soldier safety and survivability.