Here is a 72 autonomous racing drone that uses efficient algorithms for onboard vision, state estimation and control in order to fly through a narrow racing track with a speed of 2 m/s. The drone relies on model predictions that are corrected with gate detection. Here is how the researchers explain this approach:
a strategy for autonomous drone racing which is computationally more efficient than navigation methods like visual inertial odometry and simultaneous localization and mapping. This fast light-weight vision-based navigation algorithm estimates the position of the drone by fusing race gate detections with model dynamics predictions.