MonoRace drone in Abu Dhabi outruns three former world champions — Live Science

The autonomous MonoRace drone, created by researchers at Delft University of Technology in the Netherlands, recorded a time faster than the results of three former world champions in drone racing at the 2025 Abu Dhabi Autonomous Racing League competition. The system also defeated all other autonomous teams, Live Science reports.

Drone speed and specifications

On the track in Abu Dhabi, MonoRace completed two laps through 11 gates in 16.56 seconds. Its maximum speed was approximately 102 km/h. The quadcopter weighs 966 grams and measures about 30 by 28 by 10 centimeters.

The system operated without GPS, external motion-tracking systems, stereo cameras, or a preloaded track map. For navigation, the drone uses a single front-facing line-scanning camera, an inertial measurement unit to determine acceleration and rotation, and a compact onboard computer.

More current news is available on the UA.News Telegram channel Telegram.

Training in a simulator

Sensor data are processed by a neural network that recognizes racing gates, estimates the drone’s position and movement, and generates commands to control the four propellers. The flight controller transmits them 500 times per second.

The model was trained using reinforcement learning in a physics simulator, where the system made millions of attempts. It first learned to avoid collisions and reach the first gate, and later to complete most of the track at a higher speed. During training, the researchers changed the parameters of the devices, including weight, voltage, aerodynamic drag, sensor noise, and propeller condition, so that the model could operate under different conditions.

The results are described in a study published on October 14 in the journal Science Robotics. Study co-author Guido de Croon noted that such a network can be configured not only for the minimum lap time, but also to reduce the energy consumption of robotic systems.