New MIT algorithms will allow drones to fly much faster and not crash on obstacles (+ video)

Drone racing – A relatively new sport in which drones are raced along a track with obstacles that

should be avoided at maximum speed.While drone racing is purely for entertainment, the technology that allows drones to avoid obstacles in these races could serve drones in situations where they are used for mission-critical and time-sensitive operations such as search and rescue.

The Massachusetts Institute of Technology (MIT) is trying to train drones to fly faster while avoiding obstacles in their path.

Aerospace engineersdepartments at MIT have developed an algorithm that allows drones to choose the fastest route to avoid obstacles. The algorithm combines simulations of a drone flying through a virtual obstacle course and experiments with a real drone flying through the same obstacle course in the real world.

When drones are trained with a new algorithm, they can overcome obstacle courses 20% faster than drones trained with a conventional flight path planning algorithm.

Although the MIT team found that their algorithmcan lead to significantly faster completion of the course, they also found that drones trained with their new algorithm were “intentionally” inferior to drones trained using conventional algorithms.

In professional motorsport, pilots and driversknow that sometimes you have to slow down in some areas in order to go faster in others. MIT's algorithm allows it to determine whether a drone would be better off slowing down even if competitors overtake it, in order to end up being faster.

The researchers believe that thealgorithms – this is an important step towards ensuring that future drones can quickly situation navigate. For example, this technology could one day help drones used for search and rescue missions navigate crowded and dangerous environments quickly and accurately.

Source: mit

</ p>