4 ms·

I don't think this could beat a serious human competitor. Random example of the upper end of human skill: https://www.youtube.com/watch?v=eg1r-qJ117M https://w

• Dr_Mike 5y ago
  On page 32 of the paper they provide rankings of the two pilots against whom the automated system flew (Michael Isler and Timothy Trowbridge). My personal opinion of course, but I would call both of these pilots "serious human competitors" given both have been competing since 2017 in international events and received many podium finishes. But that's also beside the point. This paper is about generating time-optimal trajectories through waypoints given the system's physical constraints (e.g. limitations in thrust and rotational rates). A time-optimal trajectory is a trajectory which is time optimal—meaning that no faster trajectory exists. Given that this algorithm generates the fastest possible trajectory through the waypoints given the physical constraints of the system, it would be impossible for even a "serious" human competitor to beat it.
  • djmips 5y ago
    Perhaps, this could be used in simulation training to bring up the level of the top pilots. It could illuminate where they are losing time to an ideal pass.
• PhasmaFelis 5y ago
  That's very impressive. Doesn't mean a bot can't do better. Optimizing trajectories like this is a relatively simple task for an algorithm.