One of the most important parts of autonomous drones is obstacle avoidance. Without the owner manually controlling a drone’s movement, it can only rely on its own programming to stay afloat.
One bump in the air can result in the drone crashing to the ground, usually with serious damage.
The Great Obstacle
Drone creators try to add much needed durability in drones by making it more resilient upon crashing. Using insect anatomy as inspiration, several drone designs can now ‘rebuild’ itself after an impactful crash landing. Others try a different approach – creating drones that can stabilize itself mid-flight in the event of a collision.
However, avoiding obstacles is still the best defense during flight. In this aspect, drones use a combination of programming, cameras, and sensors to monitor a drone’s surroundings. If a drone detects an obstacle using its many eyes, it can create countermeasures to avoid it entirely. Time is also a factor, as these projectiles often leave the drone with mere seconds to react. Without a fast response, the drone would come crashing down.
Into the Matrix
In the University of Zurich, they are doing just that. Using a combination of the three factors, their tests on drones have amazing results. With their formula, their drone can avoid projectiles thrown in its direction at over 70 percent. While not perfect, it is still a decent and reliable way to avoid obstacles in the air.
In the video, the researchers throw a ball into the drone’s direction. This leaves the drone with split seconds to react. As you can see, the drone avoids the ball by subtly shifting directions mid-air. The avoidance action resembles the famous scene in The Matrix where Neo dodges bullets in slow motion. While still not perfect, the project is still very much a success. A few tweaks and we are looking at a higher success rate in the near future.