Artificial intelligence defeats world champions in FPV drone piloting for the first time (video)
Swift is an artificial intelligence algorithm that managed to defeat the champions in FPV drone piloting. It was developed by engineers at the University of Zurich.
The algorithm is designed to control unmanned aerial vehicles. In 15 out of 25 races, artificial intelligence managed to beat the world champions in FPV drone racing and set the best lap time.
Swift competed with three champions in this discipline: Alex Vanover, Thomas Bitmatta and Marvin Schäpper. The algorithm was trained in a simulation that recreated a racing track.
The method used to train artificial intelligence was based on trial and error. This is called deep reinforcement learning. During the race, the neural network processed camera data to detect the gate through which the UAV was to fly.
At the same time, the second neural network collected information from the inertia sensor about the speed and orientation of the UAV in space. Based on this data, artificial intelligence gave commands to the FPV drone.
As we mentioned above, Swift AI won 15 out of 25 races, meaning that real people were stronger than the algorithm in 40% of the races. At the same time, artificial intelligence has always been faster at the start and made sharper turns.
Engineers at the University of Zurich believe that their algorithm can be applied in practice. For example, in the inspection of buildings and ships, as well as in search and rescue operations. In addition, the technology should be of interest to the military.
