Drone Alone – Autonomous drone defeats humans
The key to this success is the use of deep neural network that sends commands directly to the motors
A student team from the TU Delft (The Netherlands) has taken the first place in autonomous high speed drone racing. The event took place in Abu Dhabi and it is the first time that a drone has beaten human pilots in an international drone racing competition – a new milestone in the development of artificial intelligence. The race took place in a real wold under real world conditions. The drone only has one camera and a neural network that direct sends commend to the motors of the drone.
The neural control network was first developed by the Advanced Concepts Team at the European Space Agency (ESA) under the name of “Guidance and Control Nets”. ESA found that deep neural networks were able to mimic the outcomes of traditional algorithms, while requiring orders of magnitude less processing time. As it was hard to test whether the networks would perform well on real hardware in space, a collaboration was formed with the MAVLab at TU Delft.
“We now train the deep neural networks with reinforcement learning, a form of learning by trial and error. ”, says Christophe De Wagter, the student team lead. “This allows the drone to more closely approach the physical limits of the system. To get there, though, we had to redesign not only the training procedure for the control, but also how we can learn about the drone’s dynamics from its own onboard sensory data.”
With speeds up to 95.8 km/h the drone flights autonomously through the parcour. It is ease to see that this news developments will lead to other autonomous applications like self driving vehicles and robotics.
Find more details here
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
