Florian Achermann

Projects

List here all the projects

---

Predicting the Weather

The steep terrain in alpine environments can cause strong up- and downdrafts or high wind magnitudes due to channeling. These winds can significantly impact the flight performance of small UAVs, even potentially causing crashes if exceeding the aircrafts flight envelope such as cruise speed or climb- and sinkrate. The goal of the project was to enable real-time volumetric wind field predictions around the UAV to enable planning for safer and more efficient flight paths. We trained a fully convolutional neural network "WindSeer" predicting the 3D wind flow and expected turbulence levels solely based on the known topography and few onboard wind observations. Using RANS-CFD simulations we generated the label flows to train the network. We then evaluated the model on held back CFD data but also on real wind measurements from UAVs and static measurement masts.

---

Project Altair

A major factor restricting the utility of UAVs is the amount of energy aboard, which limits the duration of their flights. Birds face largely the same problem, but they are adept at using their vision to aid in spotting — and exploiting — opportunities for extracting extra energy from the air around them. Within the project we developed infrared (IR) sensing techniques for detecting and mapping naturally occurring atmospheric phenomena called thermals based on a heatmap on the ground for extending the flight endurance of fixed-wing UAVs. Furthermore, we trained a neural network to detect small distortions on a single grayscale image caused by temperature gradients in the air. Detecting these so called "schlieren" could be used in the future to directly observing the thermal updraft column using an simple camera onboard an UAV.

---

Avalmapper

Snow avalanches kill, on average, 22 people per year in the Swiss Alps alone. This makes them the most deadly natural hazard, responsible for more than one third of all natural hazard fatalities. The main source of information on avalanche hazard in Switzerland for authorities, safety personnel of ski resorts and traffic lines, as well as the public, is the avalanche bulletin of the WSL Institute for Snow and Avalanche Research SLF. The goal within Avalmapper is to map avalanche activities with a UAV providing more and higher resolution information to enable better avalanche bulletins. We developed active mapping algorithms together with wind aware planners to allow safe and efficient mapping of avalanches in alpine environments to allow the drones to navigate and map avalanches in these harsh environments.

---

Solar³

Within large fields the crop and soil state can vary, thus without the appropriate information this leads certain parts of the fields to be over-/under fertilized or to wasted water. Within Solar³ the goal was to develop a remote sensing platform observing fields which would enable our project partner Gamaya build a decision support system for the farmers for more sustainable smart farming. Our part of the project was to develop the remote sensing platform SenseSoar2. It is a catapult launched solar-powered fixed-wing UAV capable of flying up to 300km in 6 hours. We prioritized developing an easy to use platform which was successfully validated by our project partners operating 20 BVLOS flights in Ukraine monitoring large fields over the summer 2019.

---

BVLOS flight over Lake Neuchâtel

We demonstrated the first networked long distance beyond visual line of sight (BVLOS) flight in Switzerland with SenseSoar2, the UAV developed as part of the Solar^{3 project. With the project partners Swisscom, v2sky, and Involi we got the permission from BAZL to fly from Orges to Bellechasse over the Lake Neuchâtel and on January 6th 2019 we successfully executed the mission.}

---

Collaborative Transport with Hybrid UAVs

Multiple small drones can transport much heavier payloads than a single drone can by itself. While this concept has been extensively studies for multicopter UAVs we explored this concept for hybrid UAVs. Such drones can travel much longer distances compared to multicopter drones but pose some challenges due to the limited maneuverability in cruise flight and the transition from hover to cruise. In this project we developed safe guidance laws and control algorithms to enable collaborative payload transport with two hybrid UAVs.