Drones War

Watch on YouTube

The military was working to protect its newly constructed aid pier in Gaza, not from a direct attack or an assault by conventional forces, but from drones. The Army is using a pair of experimental new vehicles designed to track incoming drones and then either disrupt their signal or shoot them down(More details: https://www.msn.com/en-us/news/world/the-us-army-s-drone-problem-and-how-it-s-solving-it/ar-BB1n6BLm).

All of that is happening as the various military branches consider whether or not they should each implement their own individual anti-drone corps, or work together to form a unified front against the threat that drones bring to the modern battlefield. One thing that is clear is that drones are changing warfare, with over 100,000 combat drones being used each month by both sides fighting in Ukraine.

Protecting the pier in Gaza is one way to field test new anti-drone weapons systems, although even as the pier has apparently been scouted by drones, there have been no actual attacks made against it, according to the Telegraph article. So, that may not be a great preview of how anti-drone systems might perform on a battlefield, where drone swarms are becoming an increasingly effective tactic. But there is one place where American troops and their NATO allies are already facing off against swarms of tactical drones and other unmanned vehicles, and plotting ways to combat and ultimately defeat them.

Those pitched battles are taking place in highly realistic but virtualized environments. And while there are casualties, nobody ever really gets hurt.

One of the most detailed military simulations in use today is the Virtual Battlespace 4 — or VBS4 — system, which is an entire virtualized environment designed to be able to train soldiers inside virtual reality for combat anywhere in the world. The world of VBS4 is fully simulated, meaning that soldiers are able to look at and use realistic weapons systems just like they would in real life, through the use of virtual reality headsets and other sensors. The VBS4 also recently upgraded its artificial intelligence to provide more realistic adversaries, where each individual enemy combatant is able to think for themselves, as opposed to just having an overall grand strategy. There is even a recently-launched Ukrainian Battle Simulation which lets allied countries fight realistic battles inside Ukraine against Russian troops — just in case such training would ever be needed(More details: https://bisimulations.com/products/vbs4).

The VBS4 platform is created by Bohemia Interactive Simulations(BIS), and is used by over 60 nations for intense, virtual training, including the United States. Those advancements used to come every year or so, but with the evolving battlefield in Ukraine they are now being deployed at a much more rapid pace, especially if updates pertain to drones or AI.

Drones are a step-change in warfare, providing unprecedented tactical situational awareness and a new offensive capability like loitering munitions or kamikaze drones. To reflect the changing nature of warfare, we’ve been continuously adding new drone capabilities in VBS based on real-world developments observed in publicly available information.

This allows NATO forces to develop counter-drone tactics and prototype new defense technologies within the simulation. Additionally, we've incorporated advanced AI to simulate drone behavior, reducing the need for human operators in training scenarios focused on defeating drone attacks.

The Ukraine conflict highlighted limitations in traditional simulations and wargames regarding logistics and the impact of morale on operations. Simulating force-on-force engagements like tank platoons fighting is relatively easier than accurately replicating larger-scale Brigade-level operations.

Early in the Ukrainian conflict, Russian supply chain failures and stalled armored advances due to single vehicle losses demonstrate these limitations. Simulations don't necessarily require more weapons platforms, but rather improved capabilities for simulating real soldiers operating within a conventional war.

At BISim, main approach utilizes AI to simulate every soldier individually, replacing traditional roll the dice aggregation methods and currently scaling this entity-level simulation using cloud technology.

The VBS4 is primarily used for tactical training and mission rehearsal. Scenarios are designed with specific learning objectives in mind, like a company clearing a city block or a battalion practicing a convoy mission. And VBS scenarios involve human participants ranging from a few people to hundreds, so the AI needs to support those training goals. This means it must be controllable by the administrator and repeatable to facilitate training effectiveness.

While there might be future demand for more intelligent AI in VBS, we haven't witnessed that need yet. It's important to remember that military simulations encompass various types. For instance, intelligent AI could be highly valuable for high-level wargaming, potentially identifying gaps in a commander's plan or even advising forces on the most suitable strategy based on real-time situations.

BIS current AI utilizes behavior trees for controllability and repeatability, as previously mentioned. While advancements in generative AI are exciting, their application in tactical military AI remains unclear. GenAI excels at processing vast amounts of data and offering intelligent summaries, but this differs significantly from analyzing a tactical situation and developing a solution or generating autonomous AI behavior.

BIS see machine learning AI having a more immediate impact on autonomous weapon platforms, which pose a significant future threat. Drones capable of making autonomous kill decisions without human input will be challenging to counter, and NATO needs to prioritize addressing this. Simulations like VBS can provide a testing ground for the AI driving those drones to develop tactics to counter them.

In addition to VBS's military simulation capabilities, BIS also specialize in recreating real-world environments for training and mission rehearsal and planning. A recent example is BIS high-fidelity Ukrainian terrain, developed in collaboration with LuxCarta.

This Ukrainian terrain can serve as a foundation for creating new training scenarios or even replicating real-world events, such as rehearsing using past Russian assaults. This benefits both Ukrainian and NATO forces. Our partnership with LuxCarta allows us to rapidly generate new simulation terrains, reflecting real-world changes like trench digging or infrastructure destruction.

This is a game-changer for modern militaries. They can now train virtually on terrain that precisely replicates their upcoming deployment or operational zone.

BIS enthusiastic about delivering entity-level simulation capabilities at a massive scale very soon. Traditionally, simulating every entity within large formations has been a limitation of military simulations. This is why constructive simulations rely on aggregation and so-called rolling the dice for outcomes. But our high-fidelity entity-level AI is now being combined with cloud technology to simulate massive numbers of individual units. This will be a game-changer, hopefully providing more realistic outcomes in AI-vs-AI engagements at scale.