Most of our readers will be familiar with the Skydio R1 drone that flies fully autonomously. The R1 drone is quite large and packs 13 cameras and a lot of processing power in order to fly itself and avoid obstacles. Compare that to the Crazyflie 2.0 nano quadcopter that also flies autonomously but easily fits in the palm of your hand.

The Crazyflie 2.0 nano quadcopter

The Crazyflie 2.0 nano quadcopter is said to be the smallest drone that flies completely autonomously without the need for any human guidance.

Even though computer vision has improved tremendously because of machine learning and AI, it is still difficult to deploy algorithms on drones because of memory, bandwidth and power constraints, hence the size of the Skydio R1. Researchers from ETH Zurich, Switzerland and the University of Bologna, Italy, however, have developed a way to build a tiny drone that can fly itself and only uses about 94 milliWatts (0.094 W) of energy.

The flight path of the nano quadcopter is determined by DroNet, a convolutional neural network that processes incoming images from the camera at 20 frames per second. DroNet figures out the steering angle to control the direction of the drone and to calculate the probability of a collision so that the tiny device knows if it should stop or keep going. The small unmanned aerial device was trained using thousands of images taken from bicycles and cars driving in various Urban environments.

“Computation is fully on-board, from state-estimation to navigation controls. This means nano-drones are completely autonomous. This is the first time such a small quadrotor can be controlled this way, without any need of external sensing and computing. The methodology remains however almost unchanged using steering angle and the collision probability prediction [in DroNet],” Loquercio told The Register.

Unlike the Skydio R1 drone, the Crazyflie 2.0 nano quadcopter can only avoid obstacles by moving horizontally and not up or down. For now, the tiny aerial device only works in controlled environments. Loquercio says:

“[It] only works in limited experiments, where the surroundings and navigation tasks are similar to the ones in the training dataset. It won’t fly very well in, say, forests or particularly challenging weather conditions. In the future, I see them working similar to flies. Despite not [having an] elegant flying patterns – flies crash a lot – they can reach any place they need.”

With drone technology improving so rapidly and becoming smaller, you cannot help but wonder what drones will be like in five or ten years from now.

Email sign-up form

Would you like to receive our DroneRise email every weekday morning? Enter your email below and look for an activation email in your inbox to confirm your DroneRise email sign-up.

[protected-iframe id=”5578de37b426c36a4896f2ed02c92acd-135826796-118517251″ info=”” ]

Note: Support DroneDJ by buying your next drone through our site. You can use the following links directly from manufacturers, such as DJIParrotYuneec or retailers like AmazonB&HBestBuy or eBay. Thank you!

About the Author