Intelligent coordination of manned and unmanned aerial vehicles in dynamic rescue operations

AI and aerial robotics for civil security

Project description

The aim of MENTHON is to achieve interoperability among first responder organisations, specifically in enabling highly automated coordination among manned and unmanned aerial vehicles. MENTHON co-develops innovative solutions closely with Penzberg Mountain Rescue (Bergwacht Penzberg) and DGzRS (The German Maritime Search and Rescue Service).

In this project, the UniBw M works towards AI-based algorithms to support the complex decision-making process of drone operators in Search and Rescue (SAR) operations. With this, we make a step towards “swarm-”based solutions for faster and more efficient first response.

Consortium-partners:

  1. HAT.tec GmbH
  2. AID Innovations GmbH
  3. Universtät der Bundeswehr München

    

Duration: 01.10.2022—30.06.2025

 

Funded by: Managed by:
Ministerium_Logo   VDI_Logo

 

Research priorities of the project:

Multi-vehicle routing

To date, the operation of drones in first response relies substantially on manual planning, control and communication, requiring thereby multiple drone operators to fly a single drone.
Although beneficial, increasing the number of drones in SAR operations implies the need of more manpower in drone operation. This is however not always feasible or can even be very costly, due to the long mission duration.
We investigate the use of multi-vehicle routing methods and AI-based multi-agent planning to automatise the strategic planning for multiple drones in SAR operations. By doing so, we reduce the workload for the drone operators, making it possible to increase fleet size without demanding more manpower.

Onboard intelligence for drones

Current command and control for drones rely on stable drone-operator communication, which is not always possible in a disaster-struck area. Whenever communication infrastructure is not reliable, onboard intelligence for drones becomes essential.
We investigate the use of advanced Reinforcement-Learning algorithms for onboard decision-making capabilities under the considerations of
• other cooperative and non-cooperative airborne vehicles,
• mission objectives, and
• safety constraints.

Principal investigator

Research associates

Björn Döschl M.Sc.

Björn Döschl M.Sc.

Research associates
Building 41/200, Room 2213
+49 89 6004-3586