Serious research is underway to ensure that drones fly faster than ever before while avoiding obstacles with great certainty.
Drone racing is a relatively new sport: drones are flown on a track full of obstacles, where even time is an opponent. However, the faster these devices fly, the more unstable they become. Accidents are common. Of course, if drones could be made faster and more adept, they could be reliably used in off-field time-critical operations, such as searching for survivors of a natural disaster.
Aerospace engineers at the Massachusetts University of Technology (MIT) are working on this. They have developed an algorithm that allows drones to reach their destination as quickly as possible while avoiding obstacles.
Training drones to fly around obstacles is relatively easy if they are intended for slow flight. This is because the laws of aerodynamics do not need to be paid as much attention at low speeds and can be omitted from modelling the behaviour of the drone. However, these effects are much more pronounced at high speeds, and the handling of vehicles is much more difficult to predict. To understand how high-speed aerodynamics affect drones in flight, researchers need to conduct several experiments in the laboratory with drones of different speeds and trajectories to see which ones fly without crashing. All this is an expensive and often accident-inducing training process.
The MIT team took a different path: they developed a high-speed flight planning algorithm that minimizes many attempts to identify fast and safe flight paths. The algorithm combines simulations of drones flying through a virtual obstacle course with the results of drone experiments in real-world physical space.
The researchers found that the drone, trained by their algorithms, flew up to 20 per cent faster through a simple obstacle course than a drone trained with traditional design algorithms. Interestingly, the new algorithm didn’t always keep the drone ahead of its competitors on the track. In some cases, he decided to slow down the device to handle a tricky bend or save energy, eventually speeding up and overtaking his rivals.
Engineers believe that the algorithm they are developing could be an important step in enabling future drones to navigate dangerous environments full of obstacles quickly.