It is a fourteen-month competition, in the final of which drones are not allowed to use GPS and after launch they must complete the mission completely autonomously, without pilot intervention. Each of the final teams received a grant of 500 thousand euros for the development of the system. “The Czech team F4F & CTU SkyGraph showed by far the best solution in the discipline of searching for an injured person and impressed the jury with the most general solution and the greatest commercial potential,” says Kilian Seitz, the organizer of the competition from the German agency.
Each of the four final challenges was based on a situation that delivery services, rescue workers or security control rooms may encounter in the future. The drone was not given a pre-programmed route. It only received a humanly formulated instruction, from which it had to understand the meaning of the task and then decide how to complete it. The simulated scenarios that the machines solved included a customer's request that he could not come to the door, but would wave from the window of house number 14. The drone was supposed to find a person and deliver a package to the entrance directly below him. Another scenario required the drone to find an injured person in a red T-shirt and deliver a medical package to him. The bystander was lying partially hidden behind a building and leaning against a tree. In this task, the Czech team demonstrated the best solution of all participants. "I was pleased that we won this task, because many Fly4Future clients are asking for a similar system, and in addition, we recently won a large national project focused on solving the problem of fires in buildings, where finding and rescuing people is always a priority," comments Tomáš Báča, technical team leader, researcher at the Faculty of Electrical Engineering and Mechanical Engineering of the Czech Technical University and co-owner of Fly4Future.
The technological basis of the Czech system is four cameras, 3D LiDAR and a powerful on-board computer. Through them, the drone locates itself without GPS, maps the surroundings and recognizes objects necessary for the fulfillment of the assigned mission. It continuously creates a spatial map, where it inserts detected buildings, people or vehicles, the system can then "ask" the map where the object or complex situation corresponding to the human task is located, and plan further movement based on the answer. The system does not require the involvement of a robotic expert or programmer during missions of this type. Fly4Future is gradually integrating the technologies already achieved into its robotic platforms and preparing them for commercial use.
A generally usable drone controlled via natural language could help rescuers, firefighters or security dispatchers, be used in delivering parcels, in smart agriculture or in mapping areas affected by an emergency. According to the leader of the Czech team and founder of Fly4Future, Martin Saska, the participation is confirmation that the research concept is becoming a generally applicable technology. "Today, the deployment of similar robotic systems often requires a team of specialists who prepare the mission and precisely program its course. Our goal is for the user to simply describe the desired result and the system to prepare a way to achieve it itself. This significantly reduces costs and operator demands and paves the way for the practical use of flying robots in industry, logistics, security or rescue operations," he thinks. The company has long-term experience with various drone systems, and is involved, among other things, in the development of a robotic hunter with artificial intelligence together with Eagle.One and CTU, the school from which the startup originally emerged.