Neural networks enable autonomous drone flight control

What does it take for an autonomous robot to be faster around a course than human drone-pilots? At these dizzying speeds, humans need lightning-fast reflexes.

Researchers at Delft University of Technology (The Netherlands) have developed an AI system that enables a racing drone to navigate demanding courses at extreme speed using only a single camera. It reportedly outperformed even former world champion human drone-pilots in a race.

Now published in Science Robotics, the scientific work behind this development shows how compact, efficient neural-network-based control can achieve high-performance autonomous flight under real-world conditions. The underlying principles could help advance a much broader class of autonomous robots, from vehicles to mobile and humanoid systems.

Less is more

The TU Delft team set out to investigate how far artificial intelligence can push a drone with very limited sensors, onboard computing power and energy. This is a challenging problem: at high speed, even a small delay in sensing or control can mean missing a gate, colliding with an obstacle or losing control and crashing.

The result is MonoRace, an autonomous flight system developed at the MAVLab of TU Delft’s Faculty of Aerospace Engineering. Unlike previous leading autonomous racing systems, MonoRace uses just one forward-facing camera and an inertial measurement unit (IMU) for perception, navigation and control.

Seeing the track with one camera

At high speed, the drone has very little time to interpret what it sees. A neural network identifies the racing gates in the camera images, while the onboard IMU and a model of the drone estimate its position and motion. The system does all this without GPS or external motion tracking.

“Flying at very high speed with only one camera is challenging because the images are blurry, deformed and the AI system has very little time to precisely determine where the drone is,” said Stavrow Bahnam, main author of the study. “We developed a robust way of combining what the camera sees with the drone’s own motion information, allowing it to navigate the track at high speed without GPS.” The system also had to cope with imperfections that occur in real flight, including camera interference and sensor saturation.

Neural network takes direct control

The researchers also redesigned how the drone is controlled. Rather than using a conventional control system with several layers of intermediate controllers, MonoRace uses a small neural network that directly generates the commands for the drone’s motors.

The network runs onboard a tiny computing chip, the ‘flight controller’, giving control commands 500 times per second. It was trained using reinforcement learning, allowing it to learn how to control the drone while taking its physical limitations into account.

“Instead of first planning a desired flight path and conventional commands, our neural network learns to directly control the motors,” said Robin Ferede, author of the study. “This allows the controller to react extremely quickly and exploit the capabilities of the drone while remaining small enough to run onboard.”

Tested under extreme conditions

To validate the system under realistic conditions, the researchers put it to the test in a demanding international competition environment. The 2025 Abu Dhabi Autonomous Drone Racing Competition provided a particularly rigorous challenge. The drone and racing environment were designed and operated by the competition organisers rather than by the Delft researchers. The system therefore had to operate successfully on both hardware and a track that the research team did not control. Despite these constraints, the TU Delft drone reached speeds of up to 100 km/h, defeating all other autonomous teams as well as three former world champion human pilots.

“Instead of giving the robot increasingly powerful sensors and computers, we wanted to find out how much intelligence we could extract from very limited resources,” De Wagter said. “Drone racing provided an extreme environment in which to test these ideas outside the laboratory. It pushed us to develop AI that was not only fast but also robust, and able to adapt when external reference measurements were unavailable.”

Top image credit: iStock.com/ koto_feja