Unmanned aerial systems to improve safety around the Lausitz lakes

If a swimmer gets into trouble in one of the vast lakes of Lusatia, every minute counts. Often, however, there is no one on the shore to spot them in time. A research team at the Brandenburg University of Technology Cottbus-Senftenberg (BTU) has now developed a mathematical method designed to position and control autonomous drones in such a way that they can locate drowning people more quickly than any human rescue chain. The study has just been published in the journal “Optimisation and Engineering” by Springer Nature.

The problem: large lakes, limited staff

Drowning is one of the leading causes of death from unintentional injuries worldwide. According to the World Health Organisation, around 236,000 people die from it every year. The actual figure is likely to be significantly higher. In Germany, too, the number of drowning victims has risen recently: the German Life-Saving Association (DLRG) recorded 393 deaths in Germany in 2025, 85 per cent of which occurred in inland waters such as lakes, rivers and canals – often in areas that are not monitored by lifeguards at all, or only intermittently, due to a lack of staff. Furthermore, drowning usually occurs silently: those affected are rarely able to draw attention to themselves, which makes incidents in unguarded waters particularly dangerous.

The idea: hangars on the shore, life-saving from the air

As part of the ‘RescueFly’ research project, the idea of a ‘Drone-in-a-Box’ system was developed: weatherproof hangars are erected at suitable locations along the shore, housing several types of unmanned aerial vehicles. When an emergency call is received at the control centre, the drones take off automatically, head for the scene, identify the swimmer in distress using cameras and image processing, and drop a self-inflating rescue device – in addition to traditional rescue services, not as a replacement for them.

The solution: Mathematics 

To ensure this system works in an emergency, the researchers had to solve two key problems in Mathematics:

1. Where should the hangars for the unmanned aerial systems be located?

An optimisation model calculates at which points along the shore hangars need to be built in order to cover the largest possible and most heavily frequented lake area with as few locations as possible – taking into account battery life, flight time and the capacity limits of the hangars.

1. How do the drones carry out the mission?

A second model plans the flight paths of the swarm of unmanned aerial vehicles in advance – including the effects of wind, restricted zones, collision avoidance and the acceleration behaviour of the aircraft – so that several units can fly over the search area in a coordinated manner and without detours.

The field test on lakes in Lusatia

The team led by Sascha Emanuel Zell, Dr. rer. nat. Toni Schneidereit, Prof. Dr. rer. nat. Armin Fügenschuh and Prof. Dr. rer. nat. Michael Breuß has not only developed its models theoretically but has also applied them directly to a real-world area: the Lusatian Lake District around Partwitz, Geierswalde, Sedlitz and Senftenberg – former opencast mining sites that are now used as recreational lakes. To this end, the calculations incorporated, amongst other things, historical operational data from the Lusatia and East Saxony regional control centres, official landslide hazard zones from the Lusatian and Central German Mining Administration Company mbH (LMBV), as well as official exclusion zones for unmanned aerial vehicles from the Digital Platform for Unmanned Aviation (dipul). From around 940 possible locations, the model identified the most suitable hangar sites. Real-world prototypes were also tested: a water rescue drone from MINTMASTERS equipped with a ‘Restube’ rescue buoy, and the ‘dRack’ hangar system from DELTA-Fluid Industrietechnik GmbH, whose sensor and control technology was developed at Chemnitz University of Technology.

Outlook

This work is intended as the first part of a two-part series of publications. In the second part, the researchers aim to use event-driven simulation to compare the extent to which different drone configurations can reduce response times during real-life rescue operations, and to present an AI method (YOLO) for the live detection of swimmers in distress from a bird’s-eye view. Regulatory issues, such as the authorisation of unmanned aircraft systems in emergency situations, as well as costs and maintenance requirements, are deliberately not addressed and remain the subject of future research.

Publication: Zell, S. E., Schneidereit, T., Fügenschuh, A., Breuß, M. (2026): *Autonomous Unmanned Aircraft Systems for Enhanced Search and Rescue of Drowning Swimmers: Location-Allocation Optimisation and Flight Trajectory Planning.* Optimization and Engineering. Open Access.

Press contact

Kristin Ebert
Communications and Marketing
T +49 (0) 355 69-2115
kristin.ebert(at)b-tu.de