How Mathematical Models Make Drones and Rotors More Robust

A team at Munich University of Applied Sciences is using system identification to simulate rotors and drone flights. The models are designed to save energy, reduce disruptions, and increase safety.

Researchers at Munich University of Applied Sciences are working on mathematical models that can be used to predict the behavior of aircraft rotors and autonomous drones more accurately. The aim is to make technical systems more robust, energy-efficient, and less prone to failure. This involves the use of system identification, in which the relationships between the input and output variables of a system are described mathematically.

The models developed should make it possible to simulate flight behavior and control processes realistically even before physical prototypes are built. According to Sabine Wisbacher, an aerospace engineer at Munich University of Applied Sciences, the accuracy of the models also increases the safety of the systems tested and controlled with them.

Models for Rotors and Autonomous Flight Routes

In the ARCTIS research project, researchers are developing models for rotor blades for a future generation of helicopters. These are designed to be quieter and more energy-efficient thanks to an adaptive design. This requires mathematical descriptions that accurately represent the behavior of the rotor blades in order to simulate performance and loads at an early stage.

Another focus is the EndeAR project, which deals with the optimization of autonomous drone flights. Here, mathematical models are used to derive control commands during flight from current measurement data such as speed, altitude, environmental geometry, and wind conditions in order to guide the drone safely to its destination.

Grey-Box and Black-Box Approaches

Both grey-box and black-box models are used. According to Munich University of Applied Sciences, grey-box models use known physical relationships and existing algorithms, such as equations of motion or measurement results from wind tunnel and free-flight tests. This information is combined with current weather or image data to make automatic controls more robust.

Black-box models, on the other hand, are used when the relevant relationships are not yet known. In these cases, the algorithms learn from known input and output data to predict system behavior. In the case of adaptive rotor blades, this concerns, for example, the question of how integrated actuators can change the profile of a rotor blade to achieve efficiency gains.

Saving Energy and Increasing Safety

In the future, the models developed will be used not only for simulation but also for active control in real-world operations. The researchers want to further develop black-box models so that they can be used directly during flight to adjust control processes in real time.

“Thanks to the various linear mathematical models, the performance of technical systems can be improved,” says Daniel Ossmann, professor at the Faculty of Mechanical Engineering, Automotive Engineering, and Aircraft Engineering at Munich University of Applied Sciences. The projects showed that system identification can be used to develop controls that save energy and at the same time increase the safety of drones and helicopters.

This is a partly automated translation of this german article.