In practical settings, ten minutes of flight time are generally not enough for most applications. A team comprised of researchers from the University of Klagenfurt (AAU) and NASA-JPL/California is working on ways to enable the autonomous flight of drones in several stages with intermittent charging phases. Christian Brommer, AAU doctoral student, has recently published the results of his research.
When Christian Brommer came to Klagenfurt from California, he brought a subtle, barely noticeable American accent with him. We met him to discuss his recent publication on the autonomous and long-duration flight of helicopter-drones, also known as rotorcraft UAS. It quickly became clear that the mission time of such small drones, roughly the size of a shoebox is even shorter than one might generally assume.
The origins of Brommer’s work can be traced back to the time he spent at NASA’s Jet Propulsion Laboratory (JPL) at the California Institute of Technology in Pasadena. The Water & Carbon Cycles Group and the Robotic Aerial Mobility Group located there are collaborating on a particular study case: Rotorcraft UAS equipped with multispectral cameras are deployed to patrol agricultural land by air. It is hoped that data collected in this way will allow biologists to determine the condition of the plants. To date, this task has been performed by stationary instruments, which have a limited reach, and light aircraft, which lead to significantly higher costs.
Helicopter-drones can offer an agile and cost-efficient alternative, though they have a disadvantage due to their size. Equipped with a battery the size of a typical smartphone, these helicopters can only fly for a few minutes at a time.
The goal the research team initially set itself was to allow this process to run autonomously. During the flight, the helicopter-drone should recognize when the battery is running low and return to a charging station in a timely manner. Having arrived there, the on-board computer transmits the recorded data to a basestation while the helicopter battery is automatically charged. Once recharged, the drone should be able to take off again. Using this mission plan, a plot of arable land can be accurately surveyed in several stages.
For this sequence to work smoothly, several challenges must be faced, as Christian Brommer explains: “We need sophisticated state-estimation algorithms that provide an accurate location of the helicopter to navigate and touch down precisely on the one square metre charging platform. Markings applied to the charging platform help us to achieve this. These are recognized by a camera and allow greater precision for the navigation required during the landing phase. Our algorithms for the state estimation combine data from several sensors to determine the best possible position of the helicopter. While the navigation of drones in laboratory environments and the aid of motion capture systems are already very accurate, once we are outdoors, numerous factors such as wind, changes in air pressure or lighting conditions that affect the sensors make it more difficult to estimate an accurate location. Especially because the usual GPS positioning only allows an accuracy of about five meters.
The results of the study were recently presented at the internationally renowned IROS conference (IEEE/RSJ International Conference on Intelligent Robots and Systems) and in the Journal of Field Robotics (JFR), in a special issue focusing on agricultural robotics. Having first met Stephan Weiss at NASA’s research centre in California several years ago, Christian Brommer has recently joined him in Klagenfurt, to pursue his doctoral degree here. Brommer grew up in Werne, a small town in the northwest of Germany and is the first one in his family who pursues his PhD. “But life in the countryside also laid strong foundations for my subsequent move into technology and I could always count on the support of my family”, he told us during the interview. He studied at the University of Applied Sciences in Dortmund and then went to JPL to complete his Master’s degree. The initial 6-month contract turned into three and a half years, which had a lasting effect on him: “At JPL you have the opportunity to work with the very best in each discipline. The intensive concentration of knowledge in one place gives you the feeling of having many more opportunities to choose from.” Stephan Weiss, professor at the Department of Smart Systems Technologies at the University of Klagenfurt is widely regarded as the driving force in the area of state estimation for unmanned aerial vehicles (UAV). As a member of Weiss’s research group, Christian Brommer is ideally placed to take his next steps in the world of science.
Christian Brommer, Danylo Malyuta, Daniel R. Hentzen, and Roland Brockers. Long-duration autonomy for small rotorcraft UAS including recharging. In Proc. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2018. https://ieeexplore.ieee.org/document/8594111
Danylo Malyuta, Christian Brommer, Daniel Hentzen, Thomas Stastny, Roland Siegwart and Roland Brockers. Long-Duration Fully Autonomous Operation of Rotorcraft UAS for Remote-Sensing Data Acquisition. Journal of Field Robotics (JFR) Special Issue on Agricultural Robotics. DOI:10.1002/rob.21898