Autonomous Cargo Drone ALAADy Flies Using Artificial Intelligence
Aviation experts from the European Commission predict that by 2050, more and more autonomous drones will transport cargo. With its Automated Low Altitude Air Delivery (ALAADy) project, the German Institute of Flight Guidance (DLR) is researching technologically feasible, reliable, safe and at the same time economical ways to make this possible.
With a load capacity of a ton and a tare weight of two and a half tons, the autonomous cargo drone being researched is comparatively large, and flies at a height of up to 150 meters. Since February 2016, seven institutes belonging to the DLR have been working on the concept of the ALAADy drone and the complex conditions under which it will fly. The cargo drone has to fly reliably, be perfectly suited to its task, economically viable and must not endanger anyone.
Due to the safety certification required by the European Aviation Safety Agency (EASA) for the use of drones, the implications for the project were clear early on. If certification comparable to manned aircraft was the aim, it would not be worth developing such a drone. “With the costs of safety verification for development, construction, manufacture and maintenance, the business case would not be economical,” summarizes ALAADy Project Lead Johann Dauer from the DLR Institute of Flight Systems. Rather than investing in making the flight device as reliable as manned aircraft, the project team decided to minimize the risk of operating the drone. The ALAADy drone should, for example, fly lower than normal air traffic and avoid densely populated areas. The scaled prototype, a gyroplane weighing 450 kilos and with a load capacity of around 20 kilos, is currently being developed as part of a sister project, in order to substantiate the concepts with practical test flights.
Algorithms Should Ensure the Cargo Drone Flies Safely
“We have developed algorithms which allow the drone to automatically avoid anything critical and monitor whether it is in a risky situation during flight,” Dauer explains. With the help of Artificial Intelligence (AI), it should, for example, independently avoid residential areas and continuously ensure that it does not violate any of the conditions of the safety certification. Otherwise, it should abort the mission automatically and end the flight safely.
Market studies have shown that the ALADDy cargo drone has a variety of potential applications. It can reach difficult to access or inaccessible disaster areas and transport spare parts and time-critical components between industrial sites, amongst other things.
The next steps for the research project are the first operational flight of the prototype, and the analysis of the AI algorithms developed. Currently, all components are integrated into a simulation environment; ultimately the drone should be operated by a pilot via the simulator. In addition, there are questions as to whether passenger airports should be integrated into the concept, whether cargo drones can take off from the existing runways or whether they need their own, or whether they should even have their own airports. “At the end of the project, we will know whether – or with which features – the cargo drone is safe and makes sense from a cost point of view,” says Dauer.
Text by Marion Frahm