The Nexus Newsletter
Welcome to the tenth edition of The Nexus Newsletter - a bi-weekly email newsletter covering recent announcements from Applied Intuition and important news bridging national security and autonomy.
This edition of the newsletter focuses on last week’s Association of the United States Army’s (AUSA) Annual Meeting and Exposition. This was the first year that we hosted a booth at AUSA and we are already looking forward to next year’s exposition.
Continue reading for a recap of the major news stories coming out of AUSA and other updates from the intersection of national security and autonomy.
US Army eyes new ways to evaluate autonomous vehicles
Key quote: “With the RCVs, we are looking at such new systems and new approaches,” Wallace said in an interview with C4ISRNET. “A lot of the testing procedures that are currently employed are not the most efficient.”
Military combat vehicles have historically needed to pass certain benchmarks, such as a specific number of test miles driven. The robotic vehicles, which are intended to be autonomous, might not need to pass the same thresholds, he said.
Instead, Wallace said, the service could look to certify the vehicles in a virtual environment as opposed to actually driving the vehicles the same number of miles required for regular combat vehicles. Testing and evaluation metrics might also look to companies already in the autonomous vehicle space to see what could be replicated in the military sphere.
Programmers have “been very open to try new means of testing,” Wallace said. “We’re trying to draft [off] what we’ve done in the civilian world, as well, to see if we can gain efficiencies by implementing their testing standards.”
Our take:We’re excited to hear that the Army is pursuing virtual testing and evaluation (T&E) for the RCV. As Maj. Cory Wallace observed, virtual T&E will enable the RCV program to build and evaluate vehicle performance and safety against a comprehensive scenario library that covers the entire operational design domain (ODD) for the intended deployment of the RCV. In an off-road environment where there are a limitless number of possible scenarios to test - many of which are difficult or impossible to test on a test range - virtual T&E will enable the Army to systematically build coverage without relying exclusively on costly and time-consuming real-world testing. As noted in the article, virtual T&E is an industry-standard procedure in the commercial autonomous vehicle (AV) industry, and we’re happy to hear that the Army is looking to commercial AV programs for inspiration as they develop a range of autonomous systems across domains.
PAVECast: How to Test, Verify, and Validate AV Technology with Applied Intuition
Listen to the latest edition of the Partners for Automated Vehicle Education’s (PAVE) PAVEcast podcast to learn more about the importance of verification and validation (V&V) for autonomous systems! In the latest edition of PAVEcast, Kenny Hoang, Product Manager at Applied Intuition, and Sunmin Kim, Director of Public Policy at Applied Intuition, explain the importance of V&V to autonomy development, V&V best practices, and lessons from the V&V handbook that are relevant for policymakers.
PAVE Virtual Panel: “The Road to AVs: Testing and Simulation”
Autonomous systems - including those built for commercial and defense applications - require exhaustive testing to validate vehicle safety and performance against requirements.
Tomorrow, Sunmin Kim, Director of Public Policy at Applied Intuition, will join Partners for Automated Vehicle Education (PAVE) for a virtual panel discussion on testing and simulation strategies to ensure autonomous systems meet the highest standards of safety and reliability. She will discuss:
How are vehicles tested?
What do concepts like verification and validation (V&V) mean?
How do developers determine when a vehicle is ready for the road?
She will be joined by Tim Dawkins, Senior Technical Director at Einride, for the conversation.
Autonomous platooning
Earlier this year, Applied Intuition acquired Mechanical Simulation Corporation, an industry leading provider of vehicle dynamics modeling and simulation software. Accurate and realistic vehicle dynamics models are critical to creating simulations that are reflective of the real world.
Platooning, leader-follower, and other related capabilities have a number of national security and commercial applications. Platooning improves fuel economy and reduces aerodynamic drag, but also requires the near-instantaneous exchange of information between vehicles.
Watch this video from Mechanical Simulation that demonstrates how TruckSim can be used to model autonomous platooning behavior.
News we’re reading
Autonomous systems are gaining momentum in the national security space. The presence of a wide range of unmanned and autonomous systems at AUSA is only the latest evidence of that trend. Here are key quotes from articles covering AUSA, plus brief commentary from Applied Intuition’s government team:
Breaking Defense | Lighter, hybrid, & highly automated: the Army’s next-gen armor
Key quote: “Soldiers are very excited about capabilities RCV provides to help detect enemy vehicles,” Norman told reporters. “[But] soldiers still want to be in the loop for deciding and assessing the identification of those targets or those anomalies.” It’s the difference between “target detection” and “target identification,” he said.
In other words: Troops trust the RCV to pick out potential threats and targets, but they want to double-check its work rather than just accept the algorithm’s verdict on what to shoot at. One approach Norman is considering is the Army’s experimental (and controversial) ATLAS system. ATLAS automatically detects, classifies, and prioritizes targets, highlighting them on a touchscreen interface for the gunner — but the AI itself can’t open fire because it is, quite deliberately, not connected to the firing mechanism.
Likewise, Norman continued, “soldiers are really excited about the ability of RCV to execute movement tasks — to get from Point A to Point B with little human intervention — and not quite as interested at this time in having robotic combat vehicles execute actual maneuvers.” In military terms, what most of us do all our lives is merely “movement,” navigating from place to place as efficiently as possible — something artificial intelligence is getting increasingly good at, even off-road. “Maneuver” is the art of moving across terrain under threat of enemy fire, using cover and concealment to avoid destruction — a much slower and more complicated affair that soldiers aren’t yet ready to entrust to AI.
Our take: As we have discussed in this newsletter and on our blog, off-road autonomy for defense applications is notoriously difficult because vehicles must be capable of complex path planning to execute complicated maneuvers and navigate hostile environments, unlike the Point A to Point B mobility issues common in the commercial space. It makes sense that troops trust robotic combat vehicles to execute mobility challenges more than they do maneuvers - autonomy stacks continue to mature their ability to address mobility and navigation challenges in both commercial and defense markets. Proving out autonomy stack performance in addressing mobility challenges will be essential to building trust among operators. Maturing autonomous systems’ capabilities to execute complex maneuvers – and increasing warfighter trust in those systems – will be essential to the success of the RCV program in the long-term.
Department of Defense | For Northcom, Autonomous Systems May Be Key to Homeland Defense
Key quote: When Northcom requests forces through the "request for forces" process, VanHerck said, consideration is also given to the needs of other combatant commands, such as U.S. Indo-Pacific Command, U.S. European Command] and U.S. Strategic Command, for instance. There may be other needs that end up overriding what Northcom needs.
"That adds risk in the time of crisis," he said. "I think we need forces that look vastly different than what we have today, that reduce the demand for tankers, that reduce the demand for fighters. And what I'm talking about is autonomous and unmanned systems."
VanHerck said he envisioned a future where unmanned systems might be parked "off the coast" near some of the threats to the homeland, and that those same unmanned systems would have domain-awareness capabilities and might also be armed with both kinetic and non-kinetic weaponry.
"Think of [position navigation timing] denial and deception or high-power microwave or laser capabilities," he said. "Now, I'm not asking for all the fighters that are competing directly with [Indo-Pacom] or [Eucom]. And I don't need all the tankers for that, if we position those in key locations around the country. The same thing can be said for autonomous air platforms that give me domain awareness, that can loiter for 18, 24 hours and beyond, that provide domain awareness, but they also provide potentially that kinetic effect or non-kinetic effect."
Right now, VanHerck said his vision appears to be science fiction. But he said he thinks it would become possible in less than a decade.
Our take: Many of the most promising use cases for autonomous systems in a defense context involve operations in hostile environments - not only because of enemy activity or presence but because of the environments themselves. The Arctic fits the bill and promises to be a hotbed for future tension between the United States and China or Russia. Gen. VanHerck’s vision for using autonomous systems to improve Northcom’s situational awareness and fast-reaction capabilities, while reducing its reliance on forces that might be pulled to support other combatant commands in times of crisis, demonstrates a nuanced understanding of how autonomous systems can serve as a force multiplier while filling in gaps left by manned platforms.
Defense News | Army seeking AI for targeting, navigation aboard Bradley replacement
Key quote: Brig. Gen. Geoffrey Norman, the director of the Next Generation Combat Vehicles Cross-Functional Team, said both AI-enabled targeting and navigation are “absolutely essential” for the Optionally Manned Fighting Vehicle, or OMFV.
“What we’re seeking initially is well within the realm of the feasible. We’ve demonstrated those capabilities in the field in really rigorous environments,” Norman said Oct. 12 at the Association of the U.S. Army annual conference. “But as we look toward the future, and what we want to see downstream, we want to be able to accept maturing technologies.”
Our take: - A modular architecture for the vehicle’s autonomy and on-vehicle software will prove critical to the OMFV’s long-term success and utility to the warfighter. But an open architecture is no guarantee - the mission and mobility autonomy required to produce an operationally relevant combat vehicle is not the automatic result of open standards. While certain technologies have been experimentally demonstrated, a production-ready, software-enabled fighting vehicle is still a huge lift. OMFV is close to a true inflection point for how the Army acquires combat vehicles, but the program fundamentally is still taking a software on hardware approach (not truly a software-defined design), and the process of evaluating the performance and safety of autonomous systems is still uncertain. Getting the software-hardware integration and test plan right in CY23 will prove critical to the long term viability of the program. It would be refreshing to see some public discussion related to these issues, which pose real challenges to this program.
C4ISRNET | Casualty evacuations by drone, robotic mules tested by Army battle lab
Key quotes: The lab has also spent years working on a U.S. Army program — the Squad Multipurpose Equipment Transport — to reduce or remove much of the load on soldiers. That robotic mule is designed to carry soldier rucks, batteries, water and ammunition, guided either autonomously or by an offboard driver.
Soldiers with the 101st Airborne Division and 10th Mountain Division tested the vehicle, as have Marine units. But tabletop exercises earlier this year looked at how artificial intelligence could benefit the program and small unit tactical resupply, Singleton said.
Another load-lightening piece of gear featured in recent experiments is the Silent Tactical Energy Enhanced Dismount, or STEED. It’s essentially an electric military cart that can carry gear and casualties or perform other tasks that would normally require multiple soldiers.
Our take: Despite the prevalence of “armed robots” at AUSA, it’s important to acknowledge the vast multitude of use cases for autonomous systems in the defense context, including logistics and casualty evacuation. The experimentation underway by the Army’s Maneuver Battle Lab is critical to the maturation of these use cases and capabilities.
DefenseScoop | Army to explore predictive analytics and autonomous systems for Indo-Pacific logistics
Key quotes: The Indo-Pacific “is the most demanding theater” for the military currently, she noted, “because of the distances involved, and all of the, you know, obvious reasons,” she said, likely referring to the evolving conflict landscape and tensions with China. She sees room for better collaboration with industry and a need to explore how automation and predictive analytics can support such logistics and sustainment operations.
“We are doing more and more every day to use data more effectively,” Wormuth told DefenseScoop during the briefing — but she confirmed that she is not yet satisfied with how the service is using data as a strategic asset.
Our take: Logistics and sustainment are critical to the success and survivability of forces in any theater, but those missions are complicated by long travel distances and hostile or communications-denied environments. Autonomous systems will prove essential to logistics and sustainment missions in the future, allowing the joint force to commit manned platforms to the fight at hand, rather than support missions.
Thank you for reading The Nexus Newsletter. Subscribe for more news from the nexus of national security and autonomy.