SPRIND/Felix AdlerTeam HERMES from the Control of Networked Systems research group has beaten six finalists from across Europe in the 14-month innovation competition organised by the German Federal Agency for Disruptive Innovation (SPRIND) – without GPS, without a pilot, and with two drones flying simultaneously.

A drone that can find a specific house number in unfamiliar terrain without satellite navigation or human control, land next to a blue bicycle, or follow a person for ten minutes while they attempt to hide – these were among the challenges teams faced in the final of SPRIND’s ‘Fully Autonomous Flight 2.0’ competition at Erding Air Base in Bavaria at the end of August 2026. In the end, the expert jury named Team HERMES from the University of Klagenfurt the overall winner.
The competition
The German Federal Agency for Disruptive Innovation (SPRIND) launched the competition in July 2025. The aim was to develop a flight system capable of operating safely without GNSS support (GPS, Galileo, etc.) and without human intervention. Fourteen teams from research and industry across Europe entered the first stage. Following an interim assessment in January 2026, seven teams qualified for the second stage, with each team receiving funding of up to 350,000 euros. The competition had a total budget of more than 5 million euros.
The challenges
The final, held from 24 to 27 August 2026, consisted of two missions, each comprising a static and a dynamic task. Teams only saw the operational area when it was their turn to compete: they had 30 minutes to prepare, followed by 60 minutes of mission time. The specific task was not communicated via chat until the beginning of this one-hour period.
In the ‘Last-Mile Delivery’ mission, the drone had to identify house number 14 in a residential area containing six identical houses and land next to a blue bicycle. The organisers deliberately added false house numbers and bicycles in different colours to make the task more challenging. In the second task, the drone had to locate a woman waving from a window of house 15 and land there. Smoke was also introduced to interfere with the sensors.
In the ‘Search and Rescue’ mission, the drone had to search a large, unfamiliar area containing buildings and dense woodland, locate a person lying on the ground wearing a red T-shirt, and land beside them. The final challenge was to follow a person for ten minutes based on a distinguishing physical characteristic – without knowing that the person would first walk, then run, get on a bicycle, and finally hide under an umbrella.
How the team prevailed
HERMES performed particularly strongly in the tracking challenge. The drone remained with the target for several minutes while avoiding trees, buildings, and other obstacles. The team also performed consistently well in the other tasks.
The drone developed in Klagenfurt was one of the smallest deployed in the competition. While some competitors used aircraft measuring almost two metres in diameter, weighing more than 20 kilograms, and equipped with numerous expensive sensors, HERMES relied on an inertial measurement unit, a rotating 360-degree LiDAR, and two forward-facing cameras. Instead, much of the effort went into the software, which makes all decisions directly on board the drone. According to SPRIND, the system almost appeared to have a personality of its own: within fractions of a second, the drone determines whether a gap between two trees is wide enough to fly through and selects the safest route without hesitation.
HERMES was also the only team to fly two drones simultaneously. The drones communicated directly with each other, divided the tasks and search area between them, and avoided one another during several encounters in the air – demonstrating that the system can also cope with moving obstacles.
Not everything went according to plan. As the drone is a research system rather than a finished product, it was not protected against rain. When the weather changed in Erding, the team had to improvise on site.
“We made an early decision not to simply add more sensors, but to make the drone genuinely more intelligent. The fact that such a lean system was able to keep pace with the larger platforms – and even pull ahead of them in the most challenging task – shows that autonomy is fundamentally a software problem,” explains Luca Di Pierno, who designed the intelligent autonomy and played a leading role in developing the core and navigation software.
The team
HERMES is the result of work carried out by the Control of Networked Systems (CNS) research group in the Institute for Intelligent Systems Technologies, which has been conducting research into autonomous aerial vehicles for more than a decade. The project was led by Luca Di Pierno, Yannick Morel and group leader Stephan Weiss. Much of the work was carried out by six Bachelor’s students at the University of Klagenfurt: Jakob Schmidinger, Malte Loos, Klemens Wibmer, Nikita Smolianinov, Emīls Džarcāns and Oleksandr Muzychenko.
What was decided in a matter of seconds in Erding – which gap between trees was safe to fly through, or which drone should cover which part of the area – has a much longer history. Around two years ago, Luca Di Pierno, supported by the team, laid the foundations for how the sensors, onboard computers, and software should work together. The system was designed to be modular, with clearly defined interfaces between its individual components, allowing it to be developed beyond the competition rather than serving a single purpose.
Luca Di Pierno played a key role in shaping this architecture and developed the software that integrates the individual modules and coordinates the drone’s decisions during flight. Building on this foundation, the students further developed the individual components and spent countless hours in the field testing the system under real-world conditions. A scholarly publication is currently in preparation. The team had already achieved an international success in 2025, when it won a competition as part of an international Search and Rescue project in Huntsville, USA.
Stephan Weiss, head of the research group, emphasises: “This victory is the result of years of fundamental research into robust state estimation and navigation without GPS, and it demonstrates that this research is ready for the real world.”
Looking ahead
The team now aims to take the technology from research into real-world applications. It is currently working on securing funding to develop the remaining components and is aiming to establish a spin-off company as soon as possible. Potential applications include the autonomous delivery of emergency medication and other packages, searching for missing persons, and a wide range of other potential uses.
About SPRIND
The Federal Agency for Disruptive Innovation (SPRIND), based in Leipzig, supports radically new technologies on behalf of the German Federal Government. In its ‘Funke’ competitions, teams compete across several stages and are evaluated by an expert jury. The jury for ‘Fully Autonomous Flight 2.0’ consisted of Hugo Jammes, Peter Chudý, Sheila Beladinejad, Patrick Rose, Jörg Dittrich and Roberto Calandra.
Photos may be used free of charge for reporting on the SPRIND competition. Photo credit: SPRIND/Felix Adler

#top .avia-gallery.av-mtwlbhom-9dcc81da0520e0bcb798fcb71e22fee1 .avia-gallery-thumb a{
width:20%;
}

Der Beitrag Team from the University of Klagenfurt wins European drone competition ‘Fully Autonomous Flight 2.0’ erschien zuerst auf University of Klagenfurt.