New algorithm for autonomous drones outperforms human drone pilots (Video)

The new algorithm will be useful in services like package delivery and rescue missions


The researchers from the University of Zurich have made a new algorithm for flying a drone. The quadrotor drone was on a test flight with the new algorithm against a pair of human drone racing pilots. The test flight results showed that the new algorithm was faster than the ones which flew with the help of a human pilot. The time-optimal trajectories consider the limitations of the drone which the new algorithm takes into consideration.

In recent days the purposes of drones are on the rise. This includes searching for people in natural calamities. The limited battery life of the drone pushes it to be fast. The limited time calls for finding the best trajectory and the right acceleration. The new algorithm allows the drone to find these. Earlier autonomous systems could not reach the performance of the best human drone pilots. But the new algorithm offers excellent solutions with which they can easily outperform the human drone pilots.

The algorithm and two human pilots were made to fly along the same quad rotor around the race circuit for the test. The external cameras provided information to the algorithm and it also captured the motion of the drones. Despite being given earlier practice for human pilots on the circuit, the algorithm proved to be faster and also consistent. The researchers hope that the new algorithm will become useful in services like package delivery, inspection, searching rescue, and the like.