engleski

The Use of Deep Reinforcement Learning for Flying a Drone

Nowadays, unmanned aerial vehicles, commonly known as drones, are used for many different purposes. However, it is still a challenging task to fly a drone, which limits its potential for doing more useful things. This paper shows how to design, develop and test an ML-Agent simulation environment by using the Unity engine and the deep reinforcement learning algorithm. First, the drone model needs to be imported in a simulating environment where it should have an ability to fly, and then it should be made to fly using deep reinforcement learning. In addition, the drone can learn to perform a certain task to elaborate the benefits of this approach.

unmanned aerial vehicle ; flight control ; VTOL Tailsitter UAV ; machine learning ; deep reinforcement learning

nije evidentirano