In June 2024, the U.S. Defense Innovation Unit (DIU) and Ukraine’s Brave1 co-hosted the NATO-Ukraine Defense Innovators Forum in Krakow, Poland. Attendees included Ukrainian operators, Polish Ministry of Defense officials, drone manufacturers, and academics who gathered to discuss defense technology innovation in Ukraine, particularly related to uncrewed aerial vehicles (UAVs) and counter-UAV systems. This blog describes some of the key takeaways from the forum, including changing concepts of operation, battlefield realities, and technological aspirations and innovations. It also builds upon CSET’s previous blog from the Future of Drones in Ukraine conference held in Warsaw in November 2023.
Evolving Concepts of Operation (CONOPS) and Aspirations
The forum, building on the foundation laid a year earlier, continued to focus on the impact of small, cheap, and attritable first-person view (FPV) UAVs, as well as the opportunities and challenges in manufacturing them at scale. But the conversation around CONOPS and the participants’ aspirations have evolved in the intervening year, in part due to the large-scale deployment of UAVs by Ukrainian forces, and similar subsequent deployments by the Russian military. Key developments in the last year include:
Ukrainian and Russian forces have deployed small UAVs (maximum ground takeoff weight of <55 pounds) at a much larger scale than in the year prior, providing significant visibility along the frontlines. This enables more widespread and sustained intelligence, surveillance, and reconnaissance (ISR), allowing easier detection of enemy forces and maneuvers and reducing the time between target identification and engagement. Consequently, it is increasingly risky for forces to concentrate in preparation for large-scale assaults, which some observers viewed as a net benefit for entrenched defending forces. Participants emphasized the need for improved and more widely deployed electronic warfare (EW) capabilities to mitigate threats from Russian UAVs, an aspiration that has been voiced throughout the war. Moreover, due to high attrition rates and limited manpower, Ukrainian forces are exploring a number of initiatives including moving exposed drone operators farther behind the frontlines, using UAVs as drone interceptors, and automating defenses to reduce the number of exposed troops.
New opportunities to synthesize information gathered by diffuse FPVs, given their wide-scale deployment. These systems are typically operated independently by military companies and platoons. The ISR provided is primarily tactical, and real-time information sharing can be limited; operators often deploy FPVs within a small geographical area to locate nearby enemy forces and make tactical decisions within a closed loop. Aggregating and synthesizing this information could paint a clearer operational picture and support decision-making above the level of individual units, as well as help direct fires from artillery and other forces (akin to efforts by the U.S. 18th Airborne Corps, as well as systems such as Anduril’s Lattice and the Army’s Tactical Intelligence Targeting Access Node ground station).
Given the payload and range limitations of small UAVs, participants saw opportunities to coordinate UAVs and network them together. This includes developing webs of systems where control can be passed to different operators when small UAVs move beyond remotely operable ranges (typically within 18 miles from the operator), or coordinating many networked systems to enable simultaneous strikes (which could reduce the number of human operators per UAV and increase their lethality when deployed en masse). Participants noted that a standardized operating system, on which software could be integrated and updates could be rolled out, would be a key enabler for UAV networking and coordination. However, it was unclear from the discussions how these efforts might be implemented or how communication/coordination challenges would be overcome.
More ambitiously, some participants aspired for Ukraine to develop FPVs (and larger UAVs) with autonomous capabilities that could circumvent EW and enable “last-mile” strikes in contested environments without a human operator. This was exemplified by a two-day hackathon that ran in tandem with the conference, where participants sought to “provide automated last mile guidance for optimal navigation [of FPVs] without human intervention.” The winner of the hackathon, a team of engineers from Sensorama Lab and KEF Robotics, developed an “autonomous capability to visually identify objects of interest… with limited computational resources, and a platform appropriate for attritable missions.”
There was significant emphasis on the potential for computer vision (CV) onboard FPVs, but the discussions lacked details and it was ultimately unclear the extent to which AI is currently being used in the conflict or the timeframes in which it may start being used more broadly. Participants did not discuss the types of CV models that may be used, their performance and reliability under different constraints, the processes to fine-tune them, nor the quality of onboard cameras and processors required to run them (or whether the processors would be onboard the drones). These variables can impact the overall utility of CV-enabled autonomy, particularly if the models are intended to run onboard UAVs. For example, many FPVs are small, light, and have limited battery power, which means higher-end processors (such as those found in datacenters) cannot run onboard. These constraints make it more challenging to run higher-performing CV models that require more compute and can force developers to use less-performant models that are less reliable. Lastly, outside of the hackathon, it was sometimes unclear from the discussions whether the envisioned autonomy would be enabled by CV models or manually-coded instructions in the software (e.g., loitering over a pre-set area, autonomous flight path programming, and self-launch/self-recover).
Temporary Advantage and the Importance of Innovative Iteration
The competition to innovate and operationalize new technologies was a recurring theme throughout the conference. Ukraine can lead in a given innovation, but the advantages are often short-lived because Russian forces can readily identify and replicate them. The cycle of innovation, adoption, and counter-innovation underscores the importance of agility and foresight, especially because Ukraine is at a disadvantage in terms of manpower and fires. The following are observations from participants regarding the implications of this dynamic and the potential avenues for maintaining an innovative edge.
Participants were cautionary about easy innovations. The easier the innovation, the more likely it is that Russia will be able to replicate it. Early in the conflict, Ukraine was first to innovate with commercial drones to enable ISR and limited strikes, but within a short period of time (sometimes only a few months) Russia adopted the same types of systems. A key observation was that, due to Russia’s ability to duplicate “easy” innovations, Ukrainian operators should carefully consider how they would counter their own systems prior to deployment. This is exemplified by Ukraine’s growing need for EW capabilities following the proliferation of commercial FPVs.
Many participants agreed that Ukraine is striving for temporary advantage and that rapid iteration in innovation is critical to maintaining that advantage. They emphasized the need to understand challenges on the frontlines, and then allow small agile teams to rapidly experiment, test, and fail (akin to skunkworks projects). Exquisite systems should not be the short-term goal; it is more important to rapidly provide “good enough” systems to operators that can then be deployed, iterated upon, and improved over time.
Ukraine appears to have made some progress in this respect. As one participant noted, “there are no verticals in Ukraine;” if a drone manufacturer wants to get involved, it can “call Kyiv, they’ll tell you what they need, then you can prototype.” And unlike in the United States and Europe, “procurement in the Ukrainian MoD is fast.” There was a notable discussion where panelists lauded the ease with which drone companies can set up operations in Kyiv (e.g., through Brave1) and develop systems for the Ukrainian forces, but some industry attendees stated that they needed more feedback from users, as well as clearer explanations of the types of problems Ukrainian operators are facing.
Conclusion
Conference participants signaled cautious optimism around Ukraine’s ability to adapt CONOPS and leverage UAVs in innovative ways, but this was juxtaposed with concerns over Russia’s ability to replicate innovations, scale UAV manufacturing, and leverage its superior manpower.
Notably, many participants seemed confident about the feasibility and utility of developing UAVs with autonomous capabilities within relatively short timeframes. But overall, there appeared to be tension in four areas: (1) scaling cheap, attritable FPVs, (2) developing autonomous capabilities reliable and survivable enough to assist the warfighter in EW-saturated environments, (3) reducing the amount of Ukrainian personnel in harm’s way, and (4) innovating in ways that provide longer-term advantages that are more difficult for Russia to replicate.
It is unclear the extent to which these efforts can be achieved simultaneously and at scale. This is not to suggest that cheap, “good enough” autonomous systems cannot be developed; the hackathon did, for example, demonstrate how CV models could enable last-mile FPV guidance to a limited degree. Rather, there is skepticism around the performance and reliability of attritable autonomous systems that use cheap components (i.e., a tradeoff between performance and price), the degree to which those systems would assist operators and have an impact on the battlefield, and whether such innovations would be challenging for Russia to replicate.